Information processing method and program

Abstract

To provide a support device and the like capable of further improvements.SOLUTION: The information processing method executed by a processor of the support device includes: executing a series of natural language processing of a text written in document i (i=1 to n) to extract multiple words containing multiple synthesis condition words that represent synthesis conditions that are the conditions for processing to produce a target material from multiple starting materials; and generating data associated with at least one synthesis condition word and at least one word, where n is a natural number greater than or equal to 2.SELECTED DRAWING: Figure 17A

Description

[Technical field]

[0001] The present disclosure relates to an assistance device that assists in exploring synthesis processes. [Background technology]

[0002] Conventionally, systems for supporting the search for synthetic processes have been proposed (see, for example, Patent Documents 1 and 2). The system of Patent Document 1 automatically ranks and narrows down synthetic routes from among a plurality of synthetic routes corresponding to the respective synthetic processes given for a target compound to be synthesized, and proposes them. The system of Patent Document 2 predicts synthetic routes corresponding to the synthetic processes, and displays synthetic routes that are economically and industrially feasible from among the many predicted synthetic routes. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2010-9257 A [Patent Document 2] Japanese Patent Application Publication No. 11-53425 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the systems in the above-mentioned patent documents 1 and 2 require further improvement.

[0005] Therefore, the present disclosure provides an assistance device and the like that can achieve further improvements. [Means for solving the problem]

[0006] A support device according to an embodiment of the present disclosure includes a generation device and an analysis device, and the generation device executes (a-1) to (a-5) for i=1 to n to generate process information 1 to process information n, which are multiple pieces of process information, where i is a natural number and n is a natural number of 2 or more. In (a-1), a material word extraction unit included in the generation device extracts, from a document i, multiple material words including multiple starting material words indicating multiple starting materials and a target material word indicating a target material. In (a-2), a process word extraction unit included in the generation device extracts, from the document i, a process word i indicating a process i for generating the target material from the multiple starting materials. In (a-3), a process word extraction unit included in the generation device extracts, from the document i, a process word i indicating a process i for generating the target material from the multiple starting materials. A condition extraction unit extracts synthesis conditions i, which are conditions i of the processing i, from the document i, and in (a-4), a characteristic value extraction unit provided in the generation device extracts characteristic values ​​i related to the target material from the document i, and in (a-5), a generation unit provided in the generation device generates process information i indicating a procedure i for generating the target material from the multiple starting materials by associating the multiple material words, the processing word i, the synthesis conditions i, and the characteristic values ​​i with each other, and the analysis device includes a combination unit that generates composite process information including a part common to the multiple process information and a multiple part that differs among the multiple process information, and an output unit that outputs the composite process information.

[0007] In addition, this comprehensive or specific aspect may be realized by a system, a method, an integrated circuit, a computer program, or a recording medium such as a computer-readable CD-ROM, or may be realized by any combination of an apparatus, a system, a method, an integrated circuit, a computer program, and a recording medium.

[0008] The support device of the present disclosure can be further improved. Further advantages and effects of one aspect of the present disclosure will be apparent from the specification and drawings. Such advantages and / or effects are provided by some embodiments and features described in the specification and drawings, respectively, but not all of them are necessarily provided to obtain one or more identical features. Effect of the Invention

[0009] The present disclosure provides an assistance device and the like that can achieve further improvements. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of a hardware configuration of a composite process search support system according to the first embodiment. [Diagram 2] FIG. 2 is a block diagram showing an example of a functional configuration of the synthesis process search support system according to the first embodiment. [Figure 3A] FIG. 3A is a diagram showing dictionaries used in each of a material word extraction unit and a processing word extraction unit in the first embodiment. [Figure 3B] FIG. 3B is a diagram showing dictionaries used by each of the condition extraction unit and the characteristic value extraction unit in the first embodiment. [Figure 4] FIG. 4 is a diagram showing a dictionary used in the generating unit in the first embodiment. [Diagram 5] FIG. 5 is a diagram showing an example of a document stored in the document storage unit of the composite process generation device in the first embodiment. [Figure 6] FIG. 6 is a diagram showing an example of words, synthesis conditions, and characteristic values ​​extracted from an extraction target region of a document in the first embodiment. [Figure 7] FIG. 7 is a diagram showing examples of material words, processing words, synthesis conditions, and characteristic values ​​extracted in the first embodiment. [Figure 8]FIG. 8 is a diagram showing an example of a synthesis process generated as structured data in the first embodiment. [Figure 9] FIG. 9 is a diagram showing an example of a composite process generation screen displayed on the display unit in the first embodiment. [Figure 10] FIG. 10 is a diagram showing an example of a compositing flow displayed in the compositing flow window according to the first embodiment. [Figure 11] FIG. 11 is a diagram showing an example of a combined process generated by the generating unit and stored in the combined process storage unit according to the first embodiment. [Figure 12] FIG. 12 is a diagram showing an example of a composite process search screen displayed on the display unit in the first embodiment. [Figure 13] FIG. 13 is a diagram showing an example of a search result display screen displayed on the display unit in the first embodiment. [Figure 14A] FIG. 14A is a diagram showing an example of a synthesis process that is combined in the first embodiment. [Figure 14B] FIG. 14B is a diagram showing another example of the synthesis process to be combined in the first embodiment. [Figure 14C] FIG. 14C is a diagram showing still another example of the synthesis process to be combined in the first embodiment. [Figure 14D] FIG. 14D is a diagram showing yet another example of the synthesis process to be combined in the first embodiment. [Figure 15] FIG. 15 is a diagram for explaining the process of the composite unit in the first embodiment. In FIG. [Figure 16] FIG. 16 is a diagram showing an example of a composite screen displayed on the display unit in the first embodiment. [Figure 17A] FIG. 17A is a flowchart showing an example of the processing operation of the composite process generation device according to the first embodiment. [Figure 17B] FIG. 17B is a flowchart showing an example of the processing operation of the composite process analysis device according to the first embodiment. [Figure 18]FIG. 18 is a diagram showing an example of a composite screen displayed on a display unit according to the first modification of the first embodiment. [Figure 19] FIG. 19 is a diagram for explaining the process of the combining unit according to the second modification of the first embodiment. In FIG. [Figure 20A] FIG. 20A is a diagram showing an example of a composite screen displayed on a display unit according to Variation 2 of Embodiment 1. As shown in FIG. [Figure 20B] FIG. 20B is a diagram showing another example of the combined screen displayed on the display unit according to the second modification of the first embodiment. [Figure 21A] FIG. 21A is a diagram showing an example of a synthesis process that is combined in the third modification of the first embodiment. [Figure 21B] FIG. 21B is a diagram showing an example of a synthesis process that is combined in the third modification of the first embodiment. [Figure 22] FIG. 22 is a diagram for explaining the process of the combining unit according to the third modification of the first embodiment. In FIG. [Figure 23] FIG. 23 is a diagram showing an example of a composite screen displayed on a display unit according to the third modification of the first embodiment. [Figure 24] FIG. 24 is a diagram showing an example of a combination process search screen displayed by a display unit according to the fourth modification of the first embodiment. [Figure 25A] FIG. 25A is a diagram showing an example of a composite screen displayed on a display unit according to Variation 4 of Embodiment 1. As shown in FIG. [Figure 25B] FIG. 25B is a diagram showing an example of a composite screen displayed on the display unit according to the fourth modification of the first embodiment. [Figure 26] FIG. 26 is a diagram showing an example of composite process information including synthesis of intermediate materials. [Figure 27] FIG. 27 is a block diagram showing an example of a functional configuration of a synthesis process search support system according to the second embodiment. [Figure 28] FIG. 28 is a diagram showing an example of a composite screen displayed on a display unit in the second embodiment. [Figure 29]FIG. 29 is a diagram for explaining a method of inputting data into an input field of the composite screen according to the second embodiment. [Diagram 30] FIG. 30 is a block diagram showing an example of a functional configuration of a synthesis process search support system according to the third embodiment. [Diagram 31] FIG. 31 is a diagram illustrating an example of facility information according to the third embodiment. [Diagram 32] FIG. 32 is a diagram showing an example of a composite screen displayed on a display unit in the third embodiment. [Diagram 33] FIG. 33 is a diagram illustrating an example of a schematic configuration of a support device according to the fourth embodiment. As shown in FIG. [Diagram 34] FIG. 34 is a diagram showing an example of the contents of the paper bib in the fourth embodiment. [Diagram 35] FIG. 35 is a diagram showing an example of the contents of the thesis PDF in the fourth embodiment. [Diagram 36] FIG. 36 is a diagram showing another example of the content of the thesis PDF in the fourth embodiment. [Figure 37] FIG. 37 is a diagram illustrating an example of a detailed configuration of the support device according to the fourth embodiment. As shown in FIG. [Figure 38] FIG. 38 is a diagram showing an example of a data table held in the device database in the fourth embodiment. [Figure 39] FIG. 39 is a diagram showing an example of a data table held in the material database according to the fourth embodiment. As shown in FIG. [Diagram 40] FIG. 40 is a diagram showing an example of a data table held in the project database in the fourth embodiment. [Diagram 41] FIG. 41 is a diagram showing an example of a data table held in the unit dictionary database in the fourth embodiment. [Diagram 42] FIG. 42 is a diagram showing an example of the organization identification table and the author identification table in the fourth embodiment. [Figure 43A] FIG. 43A is a diagram showing an example of a data table stored in the customer database in the fourth embodiment. [Figure 43B] FIG. 43B is a diagram showing another example of a data table stored in the customer database in the fourth embodiment. [Figure 43C] FIG. 43C is a diagram showing another example of a data table stored in the customer database in the fourth embodiment. [Diagram 44] FIG. 44 is a diagram showing an example of an organization data table according to the fourth embodiment. [Figure 45A] FIG. 45A is a diagram showing an example of an estimation result related to a material by an estimation unit in the fourth embodiment. FIG. [Figure 45B] FIG. 45B is a diagram showing another example of the estimation result related to the material by the estimation unit in the fourth embodiment. [Figure 46] FIG. 46 is a diagram showing an example of displaying an estimation result regarding a material by an estimation unit in the fourth embodiment. [Figure 47] FIG. 47 is a diagram showing another example of the result of estimation regarding materials by the estimation unit in the fourth embodiment. In FIG. [Figure 48] FIG. 48 is a diagram showing an example of an estimation result regarding a device by the estimation unit in the fourth embodiment. In FIG. [Figure 49] FIG. 49 is a diagram showing another example of the estimation result regarding the device by the estimation unit in the fourth embodiment. In FIG. [Figure 50A] FIG. 50A is a flowchart showing an example of processing by the extraction unit in the fourth embodiment. [Figure 50B] FIG. 50B is a flowchart showing an example of the process of the estimation unit in the fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] (Findings on which this disclosure is based) To create new materials (also called new compounds), it is necessary to find a synthesis process from among a huge number of combinations of experimental conditions. Currently, the search for a synthesis process relies on the knowledge and experience of researchers. However, finding the optimal synthesis process from among a huge number of experimental conditions requires a lot of time and money, and even for experienced researchers, it is not easy to find a synthesis process for a new compound. For this reason, there are attempts to reduce the development costs of new materials by entrusting the search for synthesis processes to computers. The search for synthesis processes by computers is carried out by analyzing the relationship between material names, experimental procedures, and experimental conditions stored in a known database.

[0012] Regarding the analysis of the above-mentioned synthesis process, for example, Patent Document 1 discloses a synthetic route evaluation system including a calculation processing means including a quantum chemical calculation unit, a reaction mechanism analysis unit, and a synthetic route ranking unit, and a storage means for storing data related to the synthetic routes. This synthetic route evaluation system extracts an optimal synthetic route from a plurality of synthetic routes for a target compound to be synthesized.

[0013] Furthermore, the above-mentioned Patent Document 2 discloses a synthetic route design system that includes a means for predicting a route (i.e., synthetic route) for synthesizing a target molecule from a compound available on the market by applying reaction rules accumulated in a reaction knowledge base, a means for estimating a variable cost index serving as a guide for the cost required to produce the target molecule from a starting material for each predicted synthetic route, and a means for presenting the predicted synthetic routes to a user in ascending order of the estimated variable cost index. With this synthetic route design system, a user can easily determine which synthetic route is economically and industrially feasible among many predicted synthetic routes.

[0014] However, the systems disclosed in the above patents, ie, systems that aid in the exploration of synthetic processes, require further improvement.

[0015] In other words, the synthesis processes of interest are often described in natural language in documents such as papers, and the systems disclosed in the above patent documents cannot easily use the synthesis processes described in the documents. More specifically, many papers are published in the field of materials every year, and the latest synthesis processes are described in these papers. However, the systems disclosed in the above patent documents cannot easily use these latest synthesis processes, and therefore cannot adequately support the search for synthesis processes.

[0016] In order to solve such problems, an assistance device according to one embodiment of the present disclosure includes a generation device and an analysis device, and the generation device executes (a-1) to (a-5) with i=1 to n to generate process information 1 to process information n, which are multiple pieces of process information, where i is a natural number and n is a natural number of 2 or more, and in (a-1), a material word extraction unit included in the generation device extracts, from document i, multiple material words including multiple starting material words indicating multiple starting materials and a target material word indicating a target material, and in (a-2), a process word extraction unit included in the generation device extracts, from document i, a process word i indicating a process i for generating the target material from the multiple starting materials, and in (a-3), In the (a-4), a characteristic value extraction unit provided in the generating device extracts a synthesis condition i, which is a condition i of the process i, from the document i, and in the (a-5), a generating unit provided in the generating device generates process information i indicating a procedure i for generating the target material from the starting materials by associating the material words, the process word i, the synthesis condition i, and the characteristic value i with each other, and the analyzing device includes a compounding unit that generates compound process information including a part common to the process information and a part different from the process information, and an output unit that outputs the compound process information. For example, the process word extraction unit may extract the process word i indicating at least one of a heating process, a mixing process, a drying process, and a dissolving process. Alternatively, the condition extraction unit may extract at least one of a temperature, a time, a pressure, and a rotation speed, each of which is indicated by a numerical value and a unit, as the synthesis condition i. Each of the process information i and the composite process information can be said to be information indicating a synthesis process, which is a process for synthesizing a target material from a plurality of starting materials.

[0017] This allows multiple pieces of process information to be easily generated from document 1 to document n. In addition, since composite process information is generated and output from the multiple pieces of generated process information, when the composite process information is displayed, common parts and different parts of many pieces of process information can be displayed in an easy-to-understand manner. This allows appropriate support for searching for synthesis processes for new materials.

[0018] In addition, when multiple process information generated from the latest documents i are combined to generate composite process information, knowledge on material development can be kept up to date. In addition, by utilizing a huge number of documents, it is possible to support the search for synthesis processes that cover a wide range of materials.

[0019] In addition, when the different synthesis conditions 1 to n associated with the common processing word i in the plurality of process information are the different parts, the compound unit may generate a comprehensive synthesis condition including the different synthesis conditions 1 to n, and replace the synthesis condition i included in the process information i and associated with the common processing word i with the comprehensive synthesis condition, thereby generating the composite process information. For example, when the different synthesis conditions 1 to n indicate different numerical values ​​1 to n, the compound unit may generate a numerical range indicated by the minimum and maximum values ​​of the different numerical values ​​1 to n as the comprehensive synthesis condition in generating the comprehensive synthesis condition. In other words, when a plurality of processed words among the processed word 1 to the processed word n ​​are the same processed word, the compound unit generates a comprehensive synthesis condition including a plurality of synthesis conditions corresponding to the plurality of processed words from the synthesis condition 1 to the synthesis condition n, replaces the plurality of synthesis conditions with the comprehensive synthesis condition, and generates the composite process information. When the mutually different synthesis condition 1 to the synthesis condition n respectively indicate a numerical value 1 to a numerical value n, and the minimum numerical value among the numerical values ​​1 to n is a numerical value p, and the maximum numerical value is a numerical value q, and the numerical value p and the numerical value q satisfy 1≦p≦n, 1≦q≦n, and p≠q, the comprehensive synthesis condition may indicate that the numerical value p is the numerical value q.

[0020] As a result, synthesis conditions 1 to n used in documents 1 to n are displayed together as a comprehensive synthesis condition, so that a synthesis process for a new material can be easily searched for by looking at the comprehensive synthesis condition.

[0021] In addition, when the mutually different synthesis conditions 1 to n indicate mutually different numerical values ​​1 to n, respectively, the composite unit may, in generating the comprehensive synthesis condition, generate a plurality of subsets by performing clustering on a set including the mutually different numerical values ​​1 to n, and generate the comprehensive synthesis condition from each of the plurality of subsets.

[0022] As a result, a plurality of comprehensive synthesis conditions are generated by clustering, so that the tendency of synthesis condition 1 to synthesis condition n used in document 1 to document n can be grasped in more detail.

[0023] In addition, in (a-2), the processing word extraction unit may extract at least one processing word i including the processing word i from the document i, and in (a-3), the condition extraction unit may extract at least one synthesis condition i including the synthesis condition i from the document i, and in generating the composite process information, the composite unit may generate the composite process information by associating each of the different parts with the common part so that a material production flow branches from the common part to each of the different parts.

[0024] As a result, for example, when a common process word and a plurality of different process words are included in a plurality of process information, the composite process information is generated so that the process flow branches from the process of the common process word to the processes of the plurality of different process words. As a result, even if a plurality of different process words are included in a plurality of process information, the common process and the plurality of different processes can be displayed in an easy-to-understand manner.

[0025] In addition, when a target material word common to the plurality of process information is included in the common part and characteristic values ​​1 to n that are different from one another in the plurality of process information are included in the different parts, the composite unit, in generating the composite process information, may further generate a comprehensive characteristic value that includes the different characteristic values ​​1 to n, and replace the characteristic value i included in the process information i with the comprehensive characteristic value.

[0026] As a result, even if the characteristic values ​​1 to n of the target material are different due to different synthesis conditions 1 to n used in documents 1 to n, those characteristic values ​​1 to n are displayed collectively as a comprehensive characteristic value, so that by looking at the comprehensive characteristic value, it is possible to easily search for a synthesis process for a new material.

[0027] The analysis device may further include a synthesis condition input unit that accepts input synthesis conditions, which are processing conditions, and a characteristic value estimation unit that estimates a characteristic value corresponding to the input synthesis condition possessed by the target material of the common target material word, and the output unit may further output the characteristic value estimated by the characteristic value estimation unit, and the characteristic value estimation unit may estimate the characteristic value corresponding to the input synthesis condition based on the relationship between the synthesis condition i included in each of the multiple process information and the characteristic value i included in each of the multiple process information.

[0028] As a result, when the user sets input synthesis conditions, the characteristic values ​​of the target material are estimated and displayed, making it easy to search for synthesis processes for materials having new characteristic values, for example.

[0029] In addition, in generating the multiple pieces of process information, the generation device executes (a-1) to (a-5) with i=1 to m to generate process information 1 to process information m, where m is a natural number equal to or greater than n, and the analysis device may further include a search unit that searches for the process information 1 to the process information n from the process information 1 to process information m generated by the generation device, and the search unit may search for the process information 1 to the process information n using at least one of the multiple material words, the processing word i, the synthesis condition i, and the characteristic value i.

[0030] This makes it possible to search for process information 1 to n that have at least one of the material words, processing words, synthesis conditions, and characteristic values ​​in common, and generate composite process information.

[0031] In addition, in (a-5), the generating unit may further add bibliographic information of the document i to the generated process information i, and the searching unit may further use the bibliographic information to search the process information 1 to the process information n. For example, the bibliographic information may indicate at least one of the name of the author of the document i, the publication date, the name of the organization to which the author belongs, and the number of citations.

[0032] This makes it possible to search for process information 1 to n that have a common publication period, for example, that is included in the bibliographic information, and generate composite process information.

[0033] In addition, the analysis apparatus may further include an equipment search unit that searches for equipment that can perform processing under the comprehensive synthesis conditions generated by the combination unit by referring to equipment information that indicates a list of processing conditions that can be realized by each of a plurality of pieces of equipment, and the output unit may further output information regarding the equipment searched for by the equipment search unit.

[0034] This makes it possible to easily inform the user of facilities that can perform processing under the comprehensive synthesis conditions, and as a result, the user can easily arrange for the facilities required for material synthesis.

[0035] In addition, the multiple process information generated by the generation unit and the composite process information generated by the combination unit may each be configured to be graphable, and the output unit may display the composite process information generated by the combination unit as a graph.

[0036] In this way, since the information on a plurality of processes is structured as graphable structured data, the composite process information can also be easily generated as graphable structured data. Furthermore, since the composite process information is displayed as a graph, such as a flow chart, the user can appropriately understand the composite process information.

[0037] In addition, when a target material word in a first process information and a starting material word in a second process information among the multiple process information are the same material word and are a common part in the multiple process information, the composite process information generated by the composite unit may synthesize a material indicated by the same material word, which is the common part, as an intermediate material from the multiple starting materials indicated by the first process information, and further indicate a procedure for generating a target material indicated by the second process information from the intermediate material.

[0038] This makes it possible to generate composite process information that indicates a procedure that is not described in any of documents 1 to n. Therefore, when such composite process information is displayed, it is possible to further support the search for a new composite process.

[0039] The support device may further include an analysis device that executes (b-1) to (b-3) for i = 1 to n to generate multiple datasets, dataset 1 to dataset n, in which (b-1) extracts information regarding at least one of the author of document i and the organization to which the author belongs from document i as author-related information i, (b-2) extracts information regarding a device from document i as device-related information i, and (b-3) generates dataset i that shows the author-related information i and the device-related information i in association with each other.

[0040] As a result, data sets 1 to n are generated from documents 1 to n. Data set i indicates the author-related information i and device-related information i of document i in association with each other, so by looking at data sets 1 to n, it is easy to understand which author or organization is using which device. This can improve the efficiency of sales activities by device sales companies.

[0041] In addition, in generating the multiple datasets, the analysis device may further execute (b-4) for i=1 to n, and in (b-4), extract information regarding the type of material from the document i as material information i, and in (b-3), generate the dataset i that further associates the material information i with the author-related information i.

[0042] As a result, the data set i shows the author-related information i and material information i of the document i in association with each other, so by looking at the data sets 1 to n, it is easy to understand which author or organization is using what material. This makes it possible to improve the efficiency of sales activities by material sales companies.

[0043] In addition, in generating the multiple datasets, the analysis device may further execute (b-5) for i=1 to n, in which (b-5) extracts information regarding the amount of the material from the document i as quantity information i, and in (b-3) generates the dataset i that further associates the quantity information i with the author-related information i.

[0044] As a result, the data set i shows the author-related information i, material information i, and quantity information i of the document i in association with each other. Therefore, by looking at the data sets 1 to n, it is easy to understand which author or organization is using what material and how much of it. This can further improve the efficiency of sales activities by material sales companies.

[0045] In addition, in generating the multiple datasets, the analysis device may further perform (b-6) for i=1 to n, in which (b-6) extracts information regarding a project supporting the author or the organization from the document i as project-related information i, and in (b-3) generates the dataset i that further associates the project-related information i with the author-related information i.

[0046] As a result, the data set i shows the author-related information i and project-related information of the document i in association with each other. Therefore, by looking at the data sets 1 to n, it is easy to understand which author or organization is receiving support from which project. Therefore, it is possible to estimate the repayment capacity of the author or organization, and to further improve the efficiency of sales activities to the author or organization.

[0047] A generation device according to an embodiment of the present disclosure includes a material word extraction unit, a processing word extraction unit, a condition extraction unit, a characteristic value extraction unit, and a generation unit, and executes (a-1) to (a-5) for i=1 to n to generate process information 1 to process information n, which are multiple pieces of process information, where i is a natural number and n is a natural number of 2 or more. In (a-1), the material word extraction unit extracts, from document i, multiple material words including multiple starting material words indicating multiple starting materials and a target material word indicating a target material, and in (a-2), the processing word extraction unit extracts, from document i, , extracts a processing word i indicating a process i for generating the target material from the multiple starting materials, and in (a-3), the condition extraction unit extracts a synthesis condition i which is a condition i of the process i from the document i, in (a-4), the characteristic value extraction unit extracts a characteristic value i related to the target material from the document i, and in (a-5), the generation unit generates process information i indicating a procedure i for generating the target material from the multiple starting materials by associating the multiple material words, the processing word i, the synthesis condition i, and the characteristic value i with each other.

[0048] This allows multiple pieces of process information to be easily generated from documents 1 to n. For example, process information that is structured data that can be graphed can be easily generated from a composite process described in a document in a natural language. This allows the composite process described in the document to be presented to the user in an easily understandable manner.

[0049] In addition, an analysis apparatus according to one aspect of the present disclosure includes a combination unit that generates composite process information including a part that is common to a plurality of process information and a plurality of parts that are different among the plurality of process information, and an output unit that outputs the composite process information, wherein each of the plurality of process information is information indicating a procedure for generating a target material from a plurality of starting materials, and indicates a plurality of starting material words that indicate the plurality of starting materials and a plurality of material words that include a target material word that indicates the target material, a processing word that indicates a process for generating the target material from the plurality of starting materials, synthesis conditions that are conditions for the process, and characteristic values ​​of the target material, in association with each other.

[0050] As a result, composite process information is generated from the multiple pieces of process information generated and output, so that when the composite process information is displayed, the common parts and different parts of the multiple pieces of process information can be displayed in an easy-to-understand manner. This makes it possible to appropriately support the search for synthesis processes for new materials. Therefore, if there is a lot of the latest process information, knowledge regarding material development can be kept wide and up-to-date.

[0051] In order to solve the above problem, a synthesis process search support system according to one aspect of the present disclosure includes a synthesis process generation device that generates, for each of a plurality of documents, a synthesis process that is a processing procedure for synthesizing a target material from a plurality of starting materials based on a description of the document, and a synthesis process analysis device that analyzes the plurality of synthesis processes generated by the synthesis process generation device. The synthesis process generation device includes a material word extraction unit that extracts a plurality of material words indicating the plurality of starting materials and the target material from one extraction target document of the plurality of documents, a processing word extraction unit that extracts processing words indicating a processing applied to material synthesis using the plurality of starting materials from the extraction target document, and a synthesis process analysis device that analyzes the plurality of synthesis processes generated by the synthesis process generation device. The synthesis process analysis device includes a condition extraction unit that extracts the processing conditions from the target document as synthesis conditions, a characteristic value extraction unit that extracts characteristic values ​​of the target material from the extraction target document, and a generation unit that generates the synthesis process by associating the material words, processing words, synthesis conditions, and characteristic values ​​extracted from the extraction target document with each other. The synthesis process analysis device includes a combination unit that generates a composite synthesis process that indicates a common part in the multiple synthesis processes and collectively indicates a multiple different parts in the multiple synthesis processes by combining the multiple synthesis processes generated by the synthesis process generation device, and a display unit that displays the composite synthesis process generated by the combination unit. For example, the processing word extraction unit may extract the processing word that indicates at least one of a heating process, a mixing process, a drying process, and a dissolving process. Alternatively, the condition extraction unit may extract at least one of a temperature, a time, a pressure, and a rotation speed, each of which is indicated by a numerical value and a unit, as the synthesis condition.

[0052] This makes it possible to easily generate a synthesis process, which is a procedure for synthesizing a target material from multiple starting materials, from a document. In addition, since a composite synthesis process is generated from the multiple synthesis processes generated, it is possible to clearly display the common parts and the different parts of many synthesis processes. This makes it possible to appropriately support the search for a synthesis process for a new material.

[0053] In addition, when multiple synthesis processes generated from up-to-date documents are combined, knowledge of materials development can be kept up to date, and the use of a large number of documents can support the search for synthesis processes covering a wide range of materials.

[0054] In addition, when a plurality of different synthesis conditions associated with a common processing word in a plurality of the synthesis processes are included in the plurality of different parts, the compounding unit may generate a comprehensive synthesis condition that comprehensively indicates the plurality of different synthesis conditions, and generate the composite synthesis process by replacing a synthesis condition that is included in any one of the synthesis processes to be replaced and is associated with the common processing word with the comprehensive synthesis condition. For example, when the plurality of different synthesis conditions each indicate a plurality of different numerical values, the compounding unit may generate a numerical range indicated by a minimum value and a maximum value of the plurality of different numerical values ​​as the comprehensive synthesis condition in generating the comprehensive synthesis condition.

[0055] As a result, the synthesis conditions used in each document are displayed together as a comprehensive synthesis condition, so that by looking at the comprehensive synthesis condition, a synthesis process for a new material can be easily searched for.

[0056] In addition, when the multiple different composite conditions each indicate multiple different numerical values, the composite unit may, in generating the comprehensive composite condition, generate multiple subsets by performing clustering on a set including the multiple different numerical values, and generate the comprehensive composite condition from each of the multiple subsets.

[0057] As a result, a plurality of comprehensive synthesis conditions are generated by clustering, so that the tendency of synthesis conditions used in each document can be grasped in more detail.

[0058] In addition, the composite unit may generate the composite composite process by collectively associating each of the multiple different parts with the common part such that a processing flow branches from the common part to each of the multiple different parts.

[0059] As a result, for example, when a common processing word and a plurality of different processing words are included in a plurality of synthesis processes, a composite synthesis process is generated so that the processing flow branches from the processing of the common processing word to the processing of the plurality of different processing words. As a result, even if a plurality of different processing words are included in a plurality of synthesis processes, the common processing and the plurality of different processing can be displayed in an easy-to-understand manner.

[0060] In addition, when a material word of a target material that is common to the multiple synthesis processes is included in the common part, and multiple characteristic values ​​that are different from one another in the multiple synthesis processes are included in the multiple different parts, the composite unit, in generating the composite synthesis process, may further generate a comprehensive characteristic value that comprehensively indicates the multiple different characteristic values, and replace the characteristic value included in the synthesis process to be replaced with the comprehensive characteristic value.

[0061] As a result, even if the characteristic values ​​of the target material differ due to the different synthesis conditions used in each document, these characteristic values ​​are displayed together as a comprehensive characteristic value, making it easy to explore the synthesis process of new materials by looking at the comprehensive characteristic value.

[0062] In addition, the synthesis process analysis apparatus may further include a synthesis condition input unit that accepts input synthesis conditions, which are processing conditions, and a characteristic value estimation unit that estimates a characteristic value of the common target material according to the input synthesis conditions, and the display unit may further display the characteristic value estimated by the characteristic value estimation unit, and the characteristic value estimation unit may estimate the characteristic value according to the input synthesis conditions based on a relationship between the synthesis conditions included in each of the plurality of synthesis processes and the characteristic values ​​included in each of the plurality of synthesis processes.

[0063] As a result, when the user sets input synthesis conditions, the characteristic values ​​of the target material are estimated and displayed, making it easy to search for a synthesis process for a material having a new characteristic value, for example.

[0064] In addition, the synthesis process analysis device may further include a search unit that searches for M synthesis processes (M is an integer equal to or less than N) from the N synthesis processes (N is an integer equal to or greater than 2) generated by the synthesis process generation device, and the combination unit may combine the M synthesis processes when M is 2 or greater, and the search unit may search for the M synthesis processes using at least one of the material words, the processing words, the synthesis conditions, and the characteristic values.

[0065] This makes it possible to search for and combine M synthesis processes that have at least one of the material words, processing words, synthesis conditions, and characteristic values ​​in common.

[0066] The generating unit may further add bibliographic information of a document corresponding to each of the generated synthesis processes, and the searching unit may further use the bibliographic information to search for the M synthesis processes. For example, the bibliographic information may indicate at least one of the author name, publication date, name of the organization to which the author belongs, and the number of citations of the document corresponding to the bibliographic information.

[0067] This makes it possible to search for and combine M synthesis processes that have common information, such as publication periods, included in the bibliographic information.

[0068] In addition, the synthesis process analysis apparatus may further include an equipment search unit that searches for equipment that can perform processing under the comprehensive synthesis conditions generated by the combination unit by referring to equipment information indicating a list of processing conditions that can be realized by each of a plurality of pieces of equipment, and the display unit may further display information regarding the equipment searched for by the equipment search unit.

[0069] This makes it possible to easily inform the user of facilities that can perform processing under the comprehensive synthesis conditions, and as a result, the user can easily arrange for the facilities required for material synthesis.

[0070] In addition, the multiple synthesis processes generated by the generation unit and the composite synthesis process generated by the synthesis unit may each be configured to be graphable, and the display unit may display the composite synthesis process generated by the synthesis unit as a graph.

[0071] In this way, since a plurality of composite processes are configured as graphable structured data, a composite composite process can also be easily generated as graphable structured data. Furthermore, since the composite composite process is displayed as a graph such as a flow chart, the user can appropriately understand the composite composite process.

[0072] In addition, a synthesis process generation device according to one aspect of the present disclosure includes a material word extraction unit that extracts from a document a plurality of material words indicating a plurality of starting materials and a target material, a processing word extraction unit that extracts from the document a processing word that indicates a processing applied to material synthesis using the plurality of starting materials, a condition extraction unit that extracts from the document conditions for the processing from the document, a characteristic value extraction unit that extracts characteristic values ​​of the target material from the document, and a generation unit that generates a synthesis process, which is a processing procedure for synthesizing the target material from the plurality of starting materials, by associating the extracted plurality of material words, the processing words, the synthesis conditions, and the characteristic values ​​with each other.

[0073] This makes it possible to easily generate a synthesis process, which is a procedure for synthesizing a target material from a plurality of starting materials, from a document. For example, a synthesis process, which is structured data that can be graphed, can be easily generated from a synthesis process described in a document in natural language. This makes it possible to present the synthesis process described in the document to the user in an easily understandable manner.

[0074] Moreover, a synthesis process analysis device according to one aspect of the present disclosure includes a combination unit that generates a composite synthesis process by combining a plurality of synthesis processes, which indicates parts common to the plurality of synthesis processes and collectively indicates a plurality of parts that are different from one another in the plurality of synthesis processes, and a display unit that displays the composite synthesis process generated by the combination unit, and each of the plurality of synthesis processes is a processing procedure for synthesizing a target material from a plurality of starting materials, and indicates a plurality of material words indicating the plurality of starting materials and the target material, a processing word indicating a processing applied to material synthesis using the plurality of starting materials, synthesis conditions that are conditions for the processing, and a characteristic value of the target material, in association with one another.

[0075] As a result, a composite synthesis process is generated from the multiple synthesis processes generated, and the common parts and the different parts of many synthesis processes can be displayed in an easy-to-understand manner. This makes it possible to appropriately support the search for synthesis processes for new materials. Therefore, if there are many up-to-date synthesis processes, knowledge on material development can be kept broad and up-to-date.

[0076] These comprehensive or specific aspects may be realized as a system, a method, an integrated circuit, a computer program, or a recording medium such as a computer-readable CD-ROM, or may be realized as any combination of the system, method, integrated circuit, computer program, or recording medium.

[0077] Hereinafter, the embodiment will be specifically described with reference to the drawings.

[0078] The embodiments described below are all comprehensive or specific examples. The numerical values, shapes, materials, components, component arrangement and connection forms, steps, and order of steps shown in the following embodiments are merely examples and are not intended to limit the present disclosure. Furthermore, among the components in the following embodiments, components that are not described in an independent claim showing a top concept are described as optional components.

[0079] In addition, each drawing is a schematic diagram and is not necessarily a precise illustration. In addition, in each drawing, the same components are denoted by the same reference numerals.

[0080] (Embodiment 1) [Synthesis process search support device] FIG. 1 is a diagram illustrating an example of a hardware configuration of a composite process search support system according to the first embodiment.

[0081] The composite process exploration support system 10 in this embodiment includes a computer 310, a display unit 205, a first recording medium 331, a second recording medium 332, and an operation device 340. Note that the composite process exploration support system 10 in the present disclosure is a specific example of a support device.

[0082] The computer 310 is, for example, a personal computer, and includes a CPU (Central Processing Unit) or a processor and a memory, and executes processing for supporting the search for a synthesis process. The computer 310 in this embodiment generates a plurality of synthesis processes and analyzes these synthesis processes. The synthesis processes generated are electronic data indicating a processing procedure for synthesizing a target material from a plurality of starting materials, and are also called process information. The synthesis processes generated are, for example, structured data configured to be graphable. The starting materials are materials that are the source of the target material, and the target material is a material generated by synthesis.

[0083] The display unit 205 displays an image according to the output from the computer 310. For example, the display unit 205 displays a synthesis process generated by the computer 310 and an analysis result of a plurality of synthesis processes.

[0084] The first recording medium 331 is, for example, a hard disk, and holds a plurality of documents. These documents are electronic data, for example, computerized papers. In this embodiment, such a first recording medium 331 is connected to the computer 310 via a communication network 320 such as the Internet. However, the first recording medium 331 may be connected to the computer 310 without going through the communication network 320. Note that the first recording medium 331 may be a recording medium other than a hard disk, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), or a semiconductor memory. Also, the first recording medium 331 may be volatile or non-volatile.

[0085] The second recording medium 332 is, for example, a hard disk similar to the first recording medium 331, and has a recording capacity for holding a plurality of synthesis processes generated by the computer 310. In this embodiment, such a second recording medium 332 is connected to the computer 310. However, the second recording medium 332 may be built into the computer 310.

[0086] The operation device 340 includes, for example, a keyboard or a mouse, and when operated by a user, outputs a signal indicating the result of the operation to the computer 310.

[0087] FIG. 2 is a block diagram showing an example of a functional configuration of the synthesis process search support system according to the first embodiment.

[0088] 2, a composite process exploration support system 10 in this embodiment includes a composite process generation device 100 and a composite process analysis device 200. Note that the composite process generation device 100 and the composite process analysis device 200 in this disclosure are specific examples of a generation device and an analysis device.

[0089] The synthesis process generation device 100 generates a synthesis process which is a processing procedure for synthesizing a target material from a plurality of starting materials. Specifically, for each of a plurality of documents, the synthesis process generation device 100 generates a synthesis process which is a processing procedure for synthesizing a target material from a plurality of starting materials based on the description of the document. The document may be image data or text format data. The document may also be data in a PDF (Portable Document Format) format. The PDF format document may be data with embedded text or imaged data with no embedded text.

[0090] Such synthesis process generation device 100 includes a document selection unit 101, a document storage unit 102, a material word extraction unit 103, a processing word extraction unit 104, a condition extraction unit 105, a characteristic value extraction unit 106, and a generation unit 107. The document storage unit 102 is configured as the above-mentioned first recording medium 331. Furthermore, each of the components of synthesis process generation device 100 other than the document storage unit 102, i.e., the document selection unit 101, the material word extraction unit 103, the processing word extraction unit 104, the condition extraction unit 105, the characteristic value extraction unit 106, and the generation unit 107, are realized by a computer 310.

[0091] Combined process analysis device 200 analyzes a plurality of combined processes generated by combined process creation device 100. In analyzing a plurality of combined processes, combined process analysis device 200 in this embodiment displays the combined processes.

[0092] Such synthesis process analysis device 200 includes synthesis process storage unit 201, search condition input unit 202, search unit 203, composite unit 204, and display unit 205. Synthesis process storage unit 201 is configured as second recording medium 332 described above. In addition, of synthesis process analysis device 200, search condition input unit 202, search unit 203, and composite unit 204 are realized by computer 310.

[0093] 2, operation device 340 is omitted, but composite process generation device 100 and composite process analysis device 200 execute processing according to a signal indicating a result of a user's operation output from operation device 340. In the present embodiment, display unit 205 is provided in composite process analysis device 200, but it does not have to be provided in composite process analysis device 200. In this case, display unit 205 only needs to be connected to at least one of composite process generation device 100 and composite process analysis device 200.

[0094] [Document Storage] As described above, the document storage unit 102 is configured as the first recording medium 331 shown in FIG. 1, and stores a plurality of documents.

[0095] [Document selection section] The document selection unit 101 sequentially selects, as an extraction target document, each of the multiple documents stored in the document storage unit 102. Note that the document selection unit 101 may select the extraction target document in a predetermined order, or may select the extraction target document in response to an operation by a user.

[0096] [Ingredient word extraction section] The ingredient word extraction unit 103 extracts ingredient words from each of the multiple documents stored in the document storage unit 102, and outputs the extracted ingredient words to the generation unit 107. That is, the ingredient word extraction unit 103 extracts ingredient words indicating multiple starting ingredients and target ingredients from one extraction target document among the multiple documents. For example, the ingredient word extraction unit 103 searches the extraction target document for ingredient words written in a dictionary described below, and extracts the ingredient words from the extraction target document.

[0097] [Processing word extraction section] The process word extraction unit 104 extracts at least one process word from each of the documents stored in the document storage unit 102, and outputs the extracted at least one process word to the generation unit 107. That is, the process word extraction unit 104 extracts a process word indicating a process applied to material synthesis using a plurality of starting materials from the extraction target document. For example, the process word extraction unit 104 searches the extraction target document for a process word described in a dictionary described later, and extracts the process word from the extraction target document. Specifically, the above-mentioned process is, for example, at least one of a heating process, a mixing process, a drying process, and a dissolving process. Therefore, the process word extraction unit 104 extracts a process word indicating at least one of a heating process, a mixing process, a drying process, and a dissolving process.

[0098] [Condition extraction part] The condition extraction unit 105 extracts synthesis conditions related to the processing words from each of the documents stored in the document storage unit 102, and outputs the extracted synthesis conditions to the generation unit 107. That is, the condition extraction unit 105 extracts the above-mentioned processing conditions as synthesis conditions from the extraction target document. For example, the condition extraction unit 105 searches the extraction target document for synthesis condition words described in a dictionary described later, and extracts synthesis conditions including the synthesis condition words from the extraction target document. Specifically, the above-mentioned synthesis conditions are at least one of temperature, time, pressure, and rotation speed, each of which is indicated by a numerical value and a unit. Therefore, the condition extraction unit 105 extracts at least one of temperature, time, pressure, and rotation speed, each of which is indicated by a numerical value and a unit, as a synthesis condition.

[0099] [Characteristic value extraction section] The characteristic value extraction unit 106 extracts characteristic values ​​of the target material from each of the multiple documents stored in the document storage unit 102, and outputs the extracted characteristic values ​​to the generation unit 107. That is, the characteristic value extraction unit 106 extracts characteristic values ​​of the target material from the extraction target document. For example, the characteristic value extraction unit 106 searches the extraction target document for characteristic value words written in a dictionary described below, and extracts characteristic values ​​including the characteristic value words from the extraction target document. Specifically, the characteristic values ​​are, for example, conductivity, mobility, or energy.

[0100] [Generation part] The generating unit 107 generates a synthesis process by associating multiple ingredient words, processing words, synthesis conditions, and characteristic values ​​extracted from each of multiple documents stored in the document storage unit 102. Then, the generating unit 107 outputs the generated synthesis process to the synthesis process storage unit 201. That is, the generating unit 107 generates a synthesis process by associating multiple ingredient words, processing words, synthesis conditions, and characteristic values ​​extracted from the extraction target document with each other. For example, the generating unit 107 associates multiple ingredient words, processing words, synthesis conditions, and characteristic values ​​with each other based on the associations described in a dictionary described later. The generating unit 107 may also associate multiple ingredient words, processing words, synthesis conditions, and characteristic values ​​with each other based on the order of appearance or positional relationship of each of the multiple ingredient words, processing words, synthesis conditions, and characteristic values ​​in the document.

[0101] In this manner, the generating device, which is the synthesis process generating device 100 in this embodiment, executes the following (a-1) to (a-5) for i=1 to n to generate process information 1 to n, which are multiple pieces of process information. Note that i is a natural number, and n is a natural number of 2 or more. In (a-1), the ingredient word extracting unit 103 included in the generating device extracts multiple ingredient words including multiple starting material words indicating multiple starting materials and a target material word indicating a target material from the document i. In (a-2), the process word extracting unit 104 included in the generating device extracts a process word i indicating a process i for generating a target material from multiple starting materials from the document i. In (a-3), the condition extracting unit 105 included in the generating device extracts a synthesis condition i, which is a condition i for the process i, from the document i. In (a-4), the characteristic value extracting unit 106 included in the generating device extracts a characteristic value i related to the target material from the document i. In (a-5), a generation unit 107 provided in the generation device generates process information i indicating a procedure i for generating a target material from a plurality of starting materials by associating a plurality of material words, a processing word i, a synthesis condition i, and a characteristic value i with each other.

[0102] [Synthesis Process Storage Section] For each of a plurality of composite processes generated by composite process generation device 100, composite process storage unit 201 stores the composite process in association with bibliographic information of a document corresponding to the composite process.

[0103] [Search criteria input section] Search condition input unit 202 receives search conditions for a plurality of combining processes stored in combined process storage unit 201. For example, search condition input unit 202 receives search conditions according to a signal output from operation device 340 indicating a result of a user's operation.

[0104] [Search section] Search unit 203 searches for a composite process that matches the search condition received by search condition input unit 202 from among the multiple composite processes stored in composite process storage unit 201. That is, search unit 203 searches for n composite processes (n is an integer equal to or less than m) from m composite processes (m is an integer equal to or greater than 2) generated by composite process generation device 100. In other words, in generating multiple pieces of process information, the above-mentioned generation device executes the above-mentioned (a-1) to (a-5) with i=1 to m to generate process information 1 to process information m. m is a natural number equal to or greater than n. In this case, search unit 203 included in the analysis device searches for process information 1 to process information n from process information 1 to process information m generated by the generation device.

[0105] [Combined section] The composite unit 204 composites the n composite processes when the number of composite processes searched by the search unit 203, that is, n, is 2 or more. That is, the composite unit 204 generates a composite composite process that indicates a part common to the n composite processes and collectively indicates a plurality of parts different from each other in the n composite processes by combining the n composite processes generated by the composite process generation device 100. Note that, like the composite process, the composite composite process is electronic data indicating a processing procedure for synthesizing a target material from a plurality of starting materials, and is also called composite process information. In addition, in the present disclosure, the composite of a plurality of composite processes means generating composite process information from a plurality of process information.

[0106] [Display] The display unit 205 is configured as, for example, a liquid crystal display, a plasma display, or an organic EL (Electro-Luminescence) display. Such a display unit 205 displays the composite composite process generated by the composite unit 204. In this disclosure, the display unit 205 is a specific example of an output unit. That is, the display unit 205 in this embodiment outputs the composite composite process, i.e., the composite process information, by displaying the composite composite process, i.e., the composite process information, as an image. However, the composite process information may be output by voice, or may be output to another device as electronic data.

[0107] [dictionary] Fig. 3A is a diagram showing dictionaries used by the ingredient word extraction section 103 and the processing word extraction section 104. Fig. 3B is a diagram showing dictionaries used by the condition extraction section 105 and the characteristic value extraction section 106.

[0108] The ingredient word extraction unit 103 has an ingredient word dictionary as shown in FIG. 3A(a). The ingredient word dictionary indicates a plurality of ingredient words and the attributes of those ingredient words. The attribute of an ingredient word is a starting material or a target material. For example, the ingredient word dictionary indicates the ingredient word "Li2CO3" and the attribute of that ingredient word "starting material". In addition, the ingredient word dictionary indicates the ingredient word "Li 1.1 Mg 0.6 Ti 1.9 (PO4)3" and the attribute of that ingredient word, "target ingredient". The ingredient word extraction unit 103 searches the document for ingredient words shown in this ingredient word dictionary, and extracts those ingredient words from the document. At this time, the ingredient word extraction unit 103 specifies the attributes of the extracted ingredient words by referring to the ingredient word dictionary, and adds them to those ingredient words. The ingredient word extraction unit 103 then outputs the ingredient words with the attributes added to the generation unit 107.

[0109] The processing word extraction unit 104 has a processing word dictionary shown in (b) of FIG. 3A. The processing word dictionary indicates a plurality of processing words and attributes of the processing words. For example, the processing word dictionary indicates a processing word "heat" and an attribute "heating processing" of the processing word. The processing word dictionary also indicates a processing word "mix" and an attribute "mixing processing" of the processing word. The processing word extraction unit 104 searches a document for a processing word indicated in the processing word dictionary, and extracts the processing word from the document. At this time, the processing word extraction unit 104 specifies the attributes of the extracted processing word by referring to the processing word dictionary, and adds the attributes to the processing word. Then, the processing word extraction unit 104 outputs the processing word to which the attributes have been added to the generation unit 107.

[0110] Here, when searching for a processing word shown in the processing word dictionary from a document, the processing word extraction unit 104 may not only search for the processing word itself, but also search for at least one of an inflected form and a derived form of the processing word. Furthermore, the processing word extraction unit 104 may search for the stem of the processing word shown in the processing word dictionary, and extract a word having the stem as the processing word. For example, the processing word extraction unit 104 not only searches for the processing word "heat" shown in the processing word dictionary, but also searches for "heated" and "heating". Furthermore, the processing word extraction unit 104 not only searches for the processing word "mix" shown in the processing word dictionary, but also searches for "mixtures". Furthermore, the processing word extraction unit 104 may search for the processing word from a document regardless of the character form (e.g., uppercase, lowercase, or font) of the processing word shown in the processing word dictionary.

[0111] The condition extraction unit 105 has a condition word dictionary shown in (a) of FIG. 3B. The condition word dictionary indicates a plurality of synthesis condition words and their attributes and examples. For example, the condition word dictionary indicates a synthesis condition word "℃", its attribute "temperature" and its example "at 900℃". Furthermore, for example, the condition word dictionary indicates a synthesis condition word "K", its attribute "temperature" and its example "at 300-400K". Moreover, the condition word dictionary indicates a synthesis condition word "Room Temperature", its attribute "temperature" and its example "at room temperature".

[0112] The condition extraction unit 105 searches the document for synthesis condition words shown in the condition word dictionary, and extracts synthesis conditions including the synthesis condition words from the document. Specifically, if a numerical value is used as an example of a synthesis condition word shown in the condition word dictionary, the condition extraction unit 105 extracts a synthesis condition consisting of the synthesis condition word and numerical values ​​around the synthesis condition word from the document. For example, if "at 300-400K" is shown in the condition word dictionary as an example of the synthesis condition word "K", the condition extraction unit 105 extracts a synthesis condition consisting of the synthesis condition word "K" and the numerical value immediately before it from the document. Note that the numerical value may include a power of 10 such as "M", such as "100M".

[0113] On the other hand, if a numerical value is not used in the example of the synthesis condition word shown in the condition word dictionary, the condition extraction unit 105 extracts the synthesis condition word from the document as a synthesis condition. For example, if "at room temperature" is shown in the condition word dictionary as an example of the synthesis condition word "Room Temperature", the condition extraction unit 105 extracts the synthesis condition word "room temperature" from the document as a synthesis condition. That is, at this time, the condition extraction unit 105 does not extract a numerical value. Also, the condition extraction unit 105 may search for the synthesis condition word from the document regardless of the character form (for example, uppercase, lowercase, font, etc.) of the synthesis condition word shown in the condition word dictionary.

[0114] The condition word dictionary may indicate a range of values ​​used for synthesis condition words. In this case, the condition extraction unit 105 searches for and extracts synthesis conditions that include values ​​within that range.

[0115] Furthermore, the condition extraction unit 105 specifies the attributes of the synthesis condition words included in the extracted synthesis condition by referring to the condition word dictionary, and adds them to the synthesis condition. Then, the condition extraction unit 105 outputs the synthesis condition to which the attributes have been added to the generation unit 107.

[0116] The characteristic value extraction unit 106 has a characteristic value dictionary shown in FIG. 3B(b). The characteristic value dictionary indicates a plurality of characteristic value words. A characteristic value word is a unit for indicating a characteristic value. For example, the characteristic value dictionary includes the characteristic value words "S / cm", "cm 2 / V" and "eV", etc.

[0117] The characteristic value extraction unit 106 searches the document for a characteristic value word shown in the characteristic value dictionary, and extracts a characteristic value including the characteristic value word and a numerical value immediately preceding the characteristic value word from the document. For example, the characteristic value extraction unit 106 extracts a characteristic value consisting of the characteristic value word "S / cm" and the numerical value immediately preceding the characteristic value word from the document. Alternatively, the characteristic value extraction unit 106 extracts a characteristic value consisting of the characteristic value word "eV" and the numerical value immediately preceding the characteristic value word from the document. The condition extraction unit 105 then outputs the extracted characteristic value to the generation unit 107.

[0118] FIG. 4 is a diagram showing a dictionary used in the generating unit 107. As shown in FIG.

[0119] The generation unit 107 has a first relevance dictionary shown in (a) of FIG. 4. The first relevance dictionary indicates a plurality of processing words and material attributes related to the processing words. The material attributes are starting material or target material. For example, the first relevance dictionary indicates the processing word "mix" and the material attribute "starting material" related to the processing word. The generation unit 107 associates the processing word and material words extracted from the document with each other by referring to the first relevance dictionary shown in (a) of FIG. 4.

[0120] The generation unit 107 has a second relevance dictionary shown in (b) of FIG. 4. The second relevance dictionary indicates a plurality of processing words and attributes of synthesis conditions related to those processing words. The number of attributes of a synthesis condition related to one processing word may be more than one. For example, the second relevance dictionary indicates the processing word "heat" and attribute 1 "temperature" and attribute 2 "time" of the synthesis condition related to the processing word "heat".

[0121] The generating unit 107 refers to the second relevance dictionary shown in FIG. 4(b) to associate the processing word extracted from the document with the synthesis condition.

[0122] [Generate synthesis process] FIG. 5 is a diagram showing an example of a document stored in document storage unit 102 of composite process generation device 100. As shown in FIG.

[0123] In the document storage unit 102, for example, a plurality of electronic papers are stored as documents as shown in FIG. 5. The document selection unit 101 selects one of the plurality of documents as an extraction target document. Then, the document selection unit 101 specifies an extraction target region from the extraction target document. For example, the document selection unit 101 specifies a region in which a predetermined item name is written as the extraction target region. Alternatively, the document selection unit 101 may search for a word shown in the dictionary shown in FIG. 3A or FIG. 3B, and specify a region in which the word appears most frequently, or a region in which the word appears more frequently than a threshold, as the extraction target region. Alternatively, the extraction target region may be set in advance in the document, or may be set arbitrarily by the user.

[0124] The document selection unit 101 outputs the extraction target document, whose extraction target region is explicitly indicated, to the material word extraction unit 103, the processing word extraction unit 104, the condition extraction unit 105, and the characteristic value extraction unit 106. To explicitly indicate the extraction target region, for example, a tag or a page number and a line number may be used. That is, in the extraction target document, information indicating the start and end points of the extraction target region may be indicated by a tag or the like, and the information indicating the start and end points may be indicated by a page number and a line number or the like.

[0125] In addition, in the extraction target area shown in Figure 5, in order to easily show the content written in the extraction target area, the beginning of each sentence contained in the extraction target area is numbered in the order in which the sentence appears.

[0126] FIG. 6 is a diagram showing an example of words, synthesis conditions, and characteristic values ​​extracted from an extraction target region of a document.

[0127] For example, when the document selection unit 101 specifies the extraction target area shown in FIG. 5, the ingredient word extraction unit 103 extracts ingredient words from the extraction target area by referring to the ingredient word dictionary shown in FIG. 3A(a). For example, the ingredient word extraction unit 103 extracts ingredient words from the ingredient word dictionary shown in FIG. 1.1 Mg 0.6 Ti 1.9 Extract the material words “Li2CO3”, “MgO”, “TiO2” and “NH4H2PO4” from the starting materials.

[0128] In addition, the processing word extraction unit 104 extracts processing words from the extraction target region by referring to the processing word dictionary shown in (b) of Fig. 3A. For example, the processing word extraction unit 104 extracts processing words such as "mixtures" and "milled".

[0129] In addition, the condition extraction unit 105 extracts synthesis conditions from the extraction target region by referring to the condition word dictionary shown in (a) of Fig. 3B. For example, the condition extraction unit 105 extracts synthesis conditions such as "20h", "334K", and "6h".

[0130] In addition, the characteristic value extraction unit 106 extracts the characteristic value of the target material from the extraction target region by referring to the characteristic value dictionary shown in (b) of Fig. 3B. For example, the characteristic value extraction unit 106 extracts the characteristic value "0.00013 S / cm".

[0131] FIG. 7 is a diagram showing examples of extracted material words, processing words, synthesis conditions, and characteristic values.

[0132] For example, as shown in FIG. 7(a), the material word “Li 1.1 Mg 0.6 Ti 1.9 The material words “Li2CO3”, “MgO”, “TiO2” and “NH4H2PO4” with the attribute “starting material” are extracted.

[0133] The ingredient word extraction unit 103 may extract ingredient words that are not shown in the ingredient word dictionary in FIG. 3A(a). In this case, the ingredient word extraction unit 103 may divide the sentences of the document into words and extract ingredient words from the structure of the sentences. When an ingredient word described in a document includes a variable, the ingredient word extraction unit 103 may extract ingredient words in which a numerical value is substituted for the variable. For example, when a document includes the word "Li4Ti (5-x) La (x) O 12 [x=0, 0.05, 0.1]" is written, the material word extraction unit 103 substitutes 0, 0.05, and 0.1 for x and extracts the material word "Li4Ti5O 12 ", "Li4Ti 4.95 La 0.05 O 12 " and "Li4Ti 4.9 La 0.1 O 12 " Extract.

[0134] As shown in FIG. 7B, processing words such as "mixtures," "heating," and "milled" are extracted. The attributes of the processing words are added to these extracted processing words. As shown in FIG. 7C, synthesis conditions such as "20h," "773K," and "5h" are extracted. The attributes of the synthesis conditions are added to these extracted synthesis conditions. As shown in FIG. 7C, a characteristic value of "1.30×10 -4 "S / cm" is extracted.

[0135] Each extraction unit, such as the material word extraction unit 103, the processing word extraction unit 104, the condition extraction unit 105, and the characteristic value extraction unit 106, may perform natural language processing such as morphological analysis, syntactic analysis, and context analysis on sentences described in the document. That is, each word, synthesis condition, and characteristic value may be searched for and extracted by such natural language processing. In addition, when the document is image data, each extraction unit may convert the image data into text data by performing OCR (Optical Character Recognition / Reader) on the image data, and extract each word, etc. from the text data. Alternatively, each extraction unit may extract each word, etc. by image recognition, or may extract each word, etc. using machine learning.

[0136] FIG. 8 is a diagram showing an example of a synthesis process generated as structured data.

[0137] The generating unit 107 generates a synthesis process using the material words, processing words, synthesis conditions, and characteristic values ​​shown in Fig. 7. This synthesis process is structured data and is configured to be graphable. The generating unit 107 may use the processing words themselves extracted as shown in Fig. 7 in the synthesis process, or may use the original form of the extracted processing words, that is, the processing words shown in the processing word dictionary in Fig. 3A(b), in the synthesis process.

[0138] Specifically, the generating unit 107 assigns word IDs (e.g., W1 and W2) to the extracted material words, processing words, and synthesis conditions, respectively, as shown in (a) of Fig. 8. Note that attributes are attached to the extracted material words, processing words, and synthesis conditions, respectively.

[0139] Furthermore, the generation unit 107 selects one word ID from the multiple word IDs assigned as described above, and searches for word IDs related to the selected word ID. Then, as shown in (b) of Fig. 8, the generation unit 107 associates the selected word ID with the searched word ID, and assigns a relationship ID to the pair of associated word IDs.

[0140] For example, the generation unit 107 associates the word ID of the processing word extracted from the document with the word ID of the ingredient word by referring to the first relevance dictionary shown in (a) of FIG. 4. Specifically, the generation unit 107 selects the processing word "mixtures" extracted from the document by the processing word extraction unit 104. Next, the generation unit 107 searches the first relevance dictionary for the processing word "mix" which is the original form of the processing word "mixtures". Then, the generation unit 107 identifies "starting material" as an attribute of the material related to the processing word "mix" by referring to the first relevance dictionary. Here, the attribute of the ingredient word is added to each extracted ingredient word as shown in (a) of FIG. 8.

[0141] Therefore, the generation unit 107 associates the ingredient words to which the specified attribute "starting material" is added, among the extracted multiple ingredient words, with the selected processing word "mixtures". At this time, the generation unit 107 may associate only the ingredient words, among the multiple ingredient words each having the attribute "starting material", that are within a predetermined number of words in the forward and backward directions from the selected processing word "mixtures" in a sentence included in the document, with the selected processing word "mixtures". For example, the generation unit 107 associates the ingredient word "Li2CO3" having the attribute "starting material" with the selected processing word "mixtures". As a result, as shown in (b) of FIG. 8, the word ID "W1" of the ingredient word "Li2CO3" and the word ID "W5" of the processing word "mix", which is the original form of the processing word "mixtures", are associated. The generation unit 107 assigns the relationship ID "R1" to the pair of the word ID "W1" and the word ID "W5".

[0142] Furthermore, the generating unit 107 associates the word ID of the processing word extracted from the document with the word ID of the synthesis condition by referring to the second relevance dictionary shown in (b) of FIG. 4. Specifically, the generating unit 107 selects the processing word "mixtures" extracted from the document by the processing word extraction unit 104. Next, the generating unit 107 searches the second relevance dictionary for the processing word "mix" which is the original form of the processing word "mixtures". Then, the generating unit 107 identifies "temperature" and "time" as attributes of the synthesis condition related to the processing word "mix" by referring to the second relevance dictionary. Here, the attributes of the synthesis condition are added to each extracted synthesis condition as shown in (a) of FIG. 8.

[0143] Therefore, the generating unit 107 associates the synthesis condition to which the specified attribute "temperature" is added, among the extracted synthesis conditions, with the selected processing word "mixtures". At this time, the generating unit 107 may associate only the synthesis condition, among the multiple synthesis conditions each having the attribute "temperature", that is within a predetermined number of words before and after the selected processing word "mixtures" in a sentence included in the document, with the selected processing word "mixtures". For example, the generating unit 107 associates the synthesis condition "773K" having the attribute "temperature" with the selected processing word "mixtures". As a result, as shown in (b) of FIG. 8, the word ID "W7" of the synthesis condition "773K" and the word ID "W5" of the processing word "mix", which is the original form of the processing word "mixtures", are associated. The generating unit 107 assigns the relationship ID "R6" to the pair of the word ID "W7" and the word ID "W5". In the same manner as in this process, the generation unit 107 associates, among the extracted synthesis conditions, the synthesis condition to which the specified attribute "time" is added with the selected processing word "mixtures". As a result, as shown in FIG. 8B, the word ID "W6" of the synthesis condition "20h" is associated with the word ID "W5" of the processing word "mix" which is the original form of the processing word "mixtures". The generation unit 107 assigns the relationship ID "R5" to the pair of the word ID "W6" and the word ID "W5".

[0144] Furthermore, the generating unit 107 associates the word ID of the processed word extracted from the document with the word ID of the processed word that appears next to the processed word. Specifically, the generating unit 107 selects the processed word "mixtures" extracted from the document by the processed word extraction unit 104. Next, the generating unit 107 selects the processed word "heating" that appears next to the processed word "mixtures" in the document from among the multiple processed words extracted from the document. Then, the generating unit 107 associates the processed word "mix" that is the original form of the processed word "mixtures" with the processed word "heat" that is the original form of the processed word "heating". As a result, as shown in (b) of FIG. 8, the word ID "W5" of the processed word "mix" is associated with the word ID "W8" of the processed word "heat". The generating unit 107 assigns the relationship ID "R7" to the pair of the word ID "W5" and the word ID "W8". In the case where the processed word "milled" appears next to the processed word "heating" in the document, the generation unit 107 associates the processed word "heat", which is the original form of the processed word "heating", with the processed word "mill", which is the original form of the processed word "milled", in the same manner as in the above-mentioned process. As a result, as shown in (b) of FIG. 8, the word ID "W11" of the processed word "mill" is associated with the word ID "W8" of the processed word "heat". The generation unit 107 assigns the relationship ID "R10" to the pair of the word ID "W11" and the word ID "W8".

[0145] Furthermore, the generating unit 107 assigns a characteristic value ID to the extracted characteristic value, as shown in (c) of FIG. -4A characteristic value ID "V1" is assigned to "ionic conductivity of Li4Ti5O2 / cm". The generation unit 107 also associates such a characteristic value ID with the word ID of an ingredient word having the attribute "target material". If a document contains multiple ingredient words each having the attribute "target material" and multiple ingredient values, the generation unit 107 associates the ingredient words with the ingredient values ​​based on the order in which the ingredient words appear. For example, if a document contains "ionic conductivity of Li4Ti5O2 / cm", the generation unit 107 associates the ingredient words with the ingredient values ​​based on the order in which the ingredient words appear. 12 ,Li4Ti 0.95 La 0.05 O 12 , and Li4Ti 4.9 La 0.1 O 12 were 0.1, 0.35, and 0.56 S / cm." In this case, the generating unit 107 generates the material word "Li4Ti5O 12 " and the property value "0.1S / cm", and the material word "Li4Ti 0.95 La 0.05 O 12 " and the characteristic value "0.35 S / cm". Furthermore, the generating unit 107 associates the material word "Li4Ti 4.9 La 0.1 O 12 " and the characteristic value "0.56 S / cm".

[0146] Also, if a document describes that material A is synthesized from multiple starting materials, then material B is synthesized, and then material C is synthesized, material A and material B are intermediate materials, and material C is a target material. However, the material words of material A and material B may also be extracted as material words of the target material by the material word extraction unit 103. In this case, the document also contains multiple material words each having the attribute "target material" and multiple characteristic values. Therefore, the generation unit 107 may associate the material words with the characteristic values ​​based on the order in which the material words appear, as described above.

[0147] The generating unit 107 may associate the processing word with the material word or associate the processing word with the synthesis condition by performing natural language processing, similarly to each of the extraction units described above. The generating unit 107 may also switch the association method according to the type of language described in the document (e.g., Japanese or English). For example, when the language described in the document is English, if a verb corresponding to the processing word described in the document is followed by a noun, the generating unit 107 may associate the processing word with the material word corresponding to the noun. On the other hand, when the language described in the document is Japanese, if a noun is preceded by a verb corresponding to the processing word described in the document, the generating unit 107 may associate the processing word with the material word corresponding to the noun.

[0148] FIG. 9 is a diagram showing an example of a composite process generation screen displayed on display unit 205. As shown in FIG.

[0149] For example, in response to a signal indicating the result of a user's operation output from the operation device 340, the document selection unit 101 causes the display unit 205 to display a merge process generation screen 1a shown in FIG.

[0150] The synthesis process generation screen 1a includes a list window 1b, a bibliography window 1c, a synthesis flow window 1d, a start generation button 1e, and a save button 1f.

[0151] The list window 1b displays a list of documents stored in the document storage unit 102. The start generation button 1e is a button for instructing the start of generation of a synthesis process. For example, the user operates the operation device 340 to select a document with the file name "0001" displayed in the list window 1b. Then, the user operates the operation device 340 to select the start generation button 1e. As a result, the document selection unit 101 selects the document with the file name "0001" as an extraction target document from the multiple documents stored in the document storage unit 102. Furthermore, the document selection unit 101 outputs the extraction target document to the material word extraction unit 103, the processing word extraction unit 104, the condition extraction unit 105, and the characteristic value extraction unit 106. This executes generation of a synthesis process from the extraction target document.

[0152] The bibliographic information of the document to be extracted is displayed in the bibliographic window 1c. For example, the document selection unit 101 extracts bibliographic information from the document to be extracted and displays the bibliographic information in the bibliographic window 1c. The bibliographic information includes, for example, the name of the publication journal, the publication date, the author's name, and the title. Furthermore, if the document to be extracted is a paper, the bibliographic information may include a paper ID assigned to the paper. In this case, the file name of the document may be the paper ID.

[0153] The synthesis flow window 1d displays the generated synthesis process in a graph. Hereinafter, this graphed synthesis process will be referred to as a synthesis flow. For example, when the generating unit 107 generates the synthesis process shown in FIG. 8, the generating unit 107 displays the synthesis process as a synthesis flow in the synthesis flow window 1d. This synthesis flow shows, as a flow diagram, a processing procedure in which at least one process is performed on a plurality of starting materials to generate a target material. Furthermore, the synthesis flow shows the synthesis conditions used in each process and the characteristic values ​​of the target material.

[0154] The save button 1f is a button for saving the generated compositing process. For example, when a compositing flow is displayed in the compositing flow window 1d, the user operates the operation device 340 to select the save button 1f. As a result, the generation unit 107 acquires a signal indicating the operation result from the operation device 340, and outputs the generated compositing process to the compositing process storage unit 201 in response to the signal. As a result, the generated compositing process is stored or saved in the compositing process storage unit 201.

[0155] In the example shown in FIG. 9, a composite process is generated from a document selected by a user, but a composite process may also be automatically generated from each document stored in document storage unit 102 and stored in composite process storage unit 201.

[0156] FIG. 10 is a diagram showing an example of a compositing flow displayed in the compositing flow window 1d.

[0157] The synthesis flow includes material words of multiple starting materials, at least one processing word, synthesis conditions used to process the processing words, material words of target materials, and characteristic values ​​of the target materials. The synthesis flow shown in FIG. 10 is, for example, a flow in which the synthesis process shown in FIG. 8 is graphed. That is, in the example shown in FIG. 10, the material words of the multiple starting materials are "Li2CO3", "MgO", "TiO2" and "NH4H2PO4". Also, the at least one processing word is the processing words "mix", "heat", "mill", "dry", "press" and "sinter". Each material word of the multiple starting materials is associated with the processing word "mix" by, for example, a straight line. And, as in the example shown in FIG. 10, when the synthesis flow includes multiple processing words, the multiple processing words are associated with each other by arrows.

[0158] Each processing word is associated with a synthesis condition. For example, the processing word "mix" is associated with synthesis conditions "20h" and "773K", and the processing word "heat" is associated with synthesis conditions "5h" and "1223K". The final processing word among the multiple processing words is associated with the ingredient word of the target material. In the example shown in FIG. 10, the ingredient word of the target material is "Li 1.1 Mg 0.6 Ti 1.9 (PO4)3". In other words, the last processing word "sinter" is converted into the material word "Li 1.1 Mg 0.6 Ti 1.9 An arrow pointing to "(PO4)3" associates the last processed word with the material word of the target material.

[0159] The characteristic value of the target material is placed around the material word. For example, the characteristic value "1.30 x 10 -4 S / cm" is placed.

[0160] This synthesis flow involves the mixing and heating of the starting materials Li2CO3, MgO, TiO2, and NH4H2PO4 to produce Li 1.1 Mg 0.6 Ti 1.9 The synthesis flow shows the procedure for synthesizing the target material "Li (PO4)3". The synthesis flow also shows that the "mix" treatment is carried out at a temperature of 773K for 20 hours. 1.1 Mg 0.6 Ti 1.9 The target material's characteristic value for (PO4)3 is 1.30×10 -4 This indicates that the measured value is "S / cm".

[0161] FIG. 11 is a diagram showing an example of a combining process generated by generating unit 107 and stored in combining process storage unit 201. As shown in FIG.

[0162] When the generating unit 107 stores the synthesis process shown in FIG. 8 in the synthesis process storage unit 201, the generating unit 107 adds the bibliographic information of the document corresponding to the synthesis process to the synthesis process. That is, the generating unit 107 adds the bibliographic information of the document corresponding to the synthesis process to each of the generated synthesis processes. That is, in the above-mentioned (a-5), the generating unit 107 further adds the bibliographic information of the document i to the generated process information i. The bibliographic information includes, for example, the name of the publication journal, the publication date, and the author name. The bibliographic information may also include the title of the document or the name of the organization to which the author belongs. If the document is a paper, the bibliographic information may include a paper ID assigned to the paper, and may also include the number of citations of the paper. That is, the bibliographic information may indicate at least one of the author name, publication date, organization name to which the author belongs, and number of citations of the document corresponding to the bibliographic information.

[0163] The synthesis process to which the bibliographic information is added in this manner is stored in synthesis process storage unit 201. Note that, in Fig. 11, the synthesis process is shown in the form of a synthesis flow in order to make the synthesis process easy to understand, but the synthesis process may be stored as structured data shown in Fig. 8 instead of in the form of a synthesis flow.

[0164] In this way, synthesis process generation device 100 in this embodiment can easily generate a synthesis process, which is a processing procedure for synthesizing a target material from a plurality of starting materials, from a document. For example, a synthesis process that is structured data that can be graphed can be easily generated from a synthesis process described in a document in natural language. This makes it possible to present the synthesis process described in the document to the user in an easily understandable manner.

[0165] [Analysis of the synthesis process] FIG. 12 is a diagram showing an example of a composite process search screen displayed on display unit 205. As shown in FIG.

[0166] For example, in response to a signal indicating the result of a user's operation output from the operation device 340, the search unit 203 causes the display unit 205 to display a combination process search screen 2a shown in FIG.

[0167] The synthesis process search screen 2a includes a search condition window 2b and a search start button 2e. The search condition window 2b includes a plurality of input fields for inputting the name of the starting material (i.e., material word) and a plurality of input fields for inputting the name of the process (i.e., process word). The search condition window 2b also includes an add material button 2c for adding an input field for the starting material and an add process button 2d for adding an input field for the process. The search condition window 2b also includes an input field for inputting the name of the target material (i.e., material word), an input field for inputting the publication period, and an input field for inputting the number of citations.

[0168] When such a synthesis process search screen 2a is displayed, the user performs input processing for each input field by operating the operation device 340. For example, as shown in FIG. 12, the user inputs material words "Li2CO3", "MgO", "TiO2" and "NH4H2PO4" into the four input fields for starting materials, respectively. If the input fields are insufficient, the user operates the operation device 340 to select the material addition button 2c. This causes an input field for inputting the name of the starting material to be added and displayed. Similarly, the user inputs process words "mix", "heat", "mill", "dry", "press" and "sinter" into the six input fields for processing, respectively. If the input fields are insufficient, the user operates the operation device 340 to select the process addition button 2d. This causes an input field for inputting the name of the processing to be added and displayed.

[0169] In addition, the user can enter the material word "Li 1.1 Mg 0.6 Ti 1.9Enter "(PO4)3" in the "Publication Period" field, enter "2010 / 1 / 1" to "2019 / 12 / 31" in the "Publication Period" field, and enter "10" or more in the "Number of Citations" field.

[0170] Then, the user operates the operation device 340 to select search start button 2e. As a result, the search condition input unit 202 accepts the name of the starting material, the name of the process, the name of the target material, the publication period, and the number of citations input in the search condition window 2b as search conditions. Then, the search condition input unit 202 outputs the search conditions to the search unit 203. When the search unit 203 acquires the search conditions from the search condition input unit 202, it searches for a synthesis process that matches the search conditions from among the multiple synthesis processes stored in the synthesis process storage unit 201. Note that the search conditions may include synthesis conditions and characteristic values.

[0171] In this way, the search unit 203 searches for n synthesis processes using at least one of the material words, processing words, synthesis conditions, and characteristic values. That is, the search unit 203 searches for process information 1 to process information n using at least one of a plurality of material words, processing words i, synthesis conditions i, and characteristic values ​​i. This makes it possible to search for and combine n synthesis processes that have at least one of the material words, processing words, synthesis conditions, and characteristic values ​​in common. In addition, the search unit 203 searches for n synthesis processes further using bibliographic information. That is, the search unit 203 searches for process information 1 to process information n using bibliographic information. This makes it possible to search for and combine n synthesis processes that have common information, such as publication period, included in the bibliographic information.

[0172] FIG. 13 is a diagram showing an example of a search result display screen displayed on the display unit 205. As shown in FIG.

[0173] When the search unit 203 finishes searching for the synthesis process, it causes the display unit 205 to display a search result display screen 3a shown in FIG.

[0174] The search result display screen 3a includes a search result window 3b, a first display button 3c, and a second display button 3d. A list of synthesis processes searched for by the search unit 203 is displayed in the search result window 3b. For example, a synthesis process with the name "P0001" generated from a document with a file name or paper ID "0001" is displayed in the search result window 3b. The first display button 3c is a button for displaying the synthesis process as a synthesis flow, and the second display button 3d is a button for displaying the composite synthesis process as a composite synthesis flow. The composite synthesis flow is a flow in which the composite synthesis process is graphed.

[0175] For example, the user selects a compositing process with the name "P0001" displayed in the search result window 3b by operating the operation device 340. Then, the user selects the first display button 3c by operating the operation device 340. As a result, the search unit 203 causes the display unit 205 to display the compositing process with the name "P0001" as a compositing flow, as shown in FIG. 9 or FIG. 10.

[0176] Also, the user operates the operation device 340 to select a plurality of composite processes displayed in the search result window 3b. For example, as shown in FIG. 13, a composite process with the name "P0001", a composite process with the name "P0029", a composite process with the name "P0105", and a composite process with the name "P0256" are selected. Then, the user operates the operation device 340 to select the second display button 3d. As a result, the search unit 203 outputs the four composite processes to the composition unit 204. The composition unit 204 generates a composite composite process by composing the four composite processes, and causes the display unit 205 to display the composite composite process as a composite composite flow.

[0177] 13, a composite composite process is generated from a plurality of composite processes selected by the user, but a composite composite process may be automatically generated and displayed from a plurality of composite processes found by a search. In other words, the composite unit 204 may generate a composite composite process by composing all the composite processes shown in the search result window 3b.

[0178] Furthermore, the compounding unit 204 does not generate a compound compound process when only one compound process is selected by the user or when only one compound process is found by the search. In other words, when the search unit 203 searches for n compound processes, and n is 2 or more, the compounding unit 204 compounds the n compound processes.

[0179] Each of Fig. 14A to Fig. 14D is a diagram showing an example of a composite process. Specifically, Fig. 14A is a diagram showing, as a composite flow, a composite process generated from a document that is a paper with a paper ID of "0001". Fig. 14B is a diagram showing, as a composite flow, a composite process generated from a document that is a paper with a paper ID of "0029". Fig. 14C is a diagram showing, as a composite flow, a composite process generated from a document that is a paper with a paper ID of "0105". Fig. 14D is a diagram showing, as a composite flow, a composite process generated from a document that is a paper with a paper ID of "0256".

[0180] These four synthesis processes have common parts and multiple different parts, and in Figures 14A to 14D, the multiple different parts are indicated by bold frames.

[0181] The common parts are the parts that correspond to the search conditions. That is, the material words of the starting materials, "Li2CO3", "MgO", "TiO2" and "NH4H2PO4", the processing words, "mix", "heat", "mill", "dry", "press" and "sinter", and the material word of the target material, "Li 1.1 Mg 0.6 Ti 1.9(PO4)3" is common to the four synthesis processes. In other words, the material words of the multiple starting materials, the multiple processing words, and the material words of the target material are the same in the four synthesis processes. On the other hand, the synthesis conditions and the property values ​​of the target material are different in these four synthesis processes. For example, the synthesis conditions of time associated with the processing of the processing word "heat" are different from each other in the four synthesis processes, such as "5h", "7h", and "4h". Similarly, the synthesis conditions of temperature associated with the processing of the processing word "heat" are also different from each other in the four synthesis processes, such as "1223K", "1200K", "1800K", and "900K". Furthermore, the synthesis conditions of pressure associated with the processing of the processing word "press" are also different from each other in the four synthesis processes, such as "300MPa", "250MPa", and "320MPa". And the property value of the target material is also different from each other in the four synthesis processes, such as "1.30×10 -4 S / cm, 1.32×10 -4 S / cm, 1.31×10 -4 S / cm" and "1.35×10 -4 S / cm", which are different from each other in the four synthesis processes.

[0182] The combining unit 204 combines these four combined processes to generate a combined combined process that indicates the parts common to the four combined processes and collectively indicates a plurality of parts different from each other in the four combined processes.

[0183] FIG. 15 is a diagram for explaining the processing of the composite unit 204. As shown in FIG.

[0184] The compositing unit 204 merges parts that are different from each other in each of the four composite processes described above. For example, the compositing unit 204 generates a global composite condition of time "4-7h" by merging composite conditions of time "5h", "7h", and "4h" associated with the common processing word "heat" in each of the four composite processes. Similarly, the compositing unit 204 generates a global composite condition of temperature "900-1800K" by merging composite conditions of temperature associated with the common processing word "heat" in each of the four composite processes. Furthermore, the compositing unit 204 generates a global composite condition of pressure "250-320MPa" by merging composite conditions of pressure associated with the common processing word "press" in each of the four composite processes. Furthermore, the compositing unit 204 generates a global composite condition of pressure "250-320MPa" by merging the characteristic values ​​of each of the four composite processes. -4 -1.35×10 -4 S / cm.

[0185] In this manner, in this embodiment, when a plurality of different synthesis conditions respectively indicate a plurality of different numerical values, the compound unit 204 generates a numerical range indicated by the minimum and maximum values ​​of the plurality of different numerical values ​​as a comprehensive synthesis condition. That is, in this embodiment, when a plurality of different synthesis conditions 1 to n respectively indicate a plurality of different numerical values ​​1 to n, the compound unit 204 generates a numerical range indicated by the minimum and maximum values ​​of the plurality of different numerical values ​​1 to n as a comprehensive synthesis condition in generating a comprehensive synthesis condition. In other words, when a plurality of processed words among the processed word 1 to the processed word n ​​are the same processed word, the compound unit 204 generates a comprehensive synthesis condition including a plurality of synthesis conditions corresponding to the plurality of processed words from the synthesis condition 1 to the synthesis condition n, and replaces the plurality of synthesis conditions with the comprehensive synthesis condition to generate the composite process information. Then, in the following cases, the comprehensive synthesis condition indicates a numerical value p to a numerical value q. In this case, the mutually different synthesis conditions 1 to n respectively indicate numerical values ​​1 to n, and among the numerical values ​​1 to n, the smallest numerical value is numerical value p and the largest numerical value is numerical value q, and the numerical values ​​p and q satisfy 1≦p≦n, 1≦q≦n, and p≠q.

[0186] In addition, in this embodiment, when a plurality of different synthesis conditions associated with a common processing word in four synthesis processes are included in the above-mentioned plurality of different parts, the compound unit 204 generates a comprehensive synthesis condition that comprehensively indicates the plurality of different synthesis conditions. Then, the compound unit 204 generates a composite synthesis process by replacing a synthesis condition included in any one of the four synthesis processes to be replaced and associated with a common processing word with the comprehensive synthesis condition. That is, in this embodiment, when the different synthesis conditions 1 to n associated with a common processing word i in a plurality of process information are the above-mentioned plurality of different parts, the compound unit 204 generates a comprehensive synthesis condition that includes the different synthesis conditions 1 to n. Then, the compound unit 204 generates composite process information by replacing a synthesis condition i included in the process information i and associated with the common processing word i with the comprehensive synthesis condition.

[0187] In addition, in this embodiment, when a material word of a target material common to four synthesis processes is included in the above-mentioned common part, and a plurality of characteristic values ​​different from each other in the four synthesis processes are included in the above-mentioned plurality of different parts, the compositing unit 204 further generates a global characteristic value that collectively indicates the plurality of different characteristic values. Then, the compositing unit 204 replaces the characteristic value included in the above-mentioned synthesis process to be replaced with the global characteristic value. In other words, in this embodiment, when a material word of a target material common to a plurality of process information is included in the above-mentioned common part, and a plurality of characteristic values ​​1 to n different from each other in the plurality of process information are included in the above-mentioned plurality of different parts, the compositing unit 204 further generates a global characteristic value including the mutually different characteristic values ​​1 to n, and replaces the characteristic value i included in the process information i with the global characteristic value.

[0188] For example, the composite unit 204 selects a composite process with a paper ID of "0001" as a replacement target composite process, replaces the composite conditions included in the replacement target composite process with the comprehensive composite conditions, and replaces the property values ​​included in the replacement target composite process with the comprehensive property values. This generates a composite composite process that shows the parts common to the four composite processes and comprehensively shows multiple parts that are different from each other in the four composite processes.

[0189] FIG. 16 is a diagram showing an example of a composite screen displayed on display unit 205. As shown in FIG.

[0190] When the composite unit 204 generates the composite composite process as described above, it causes the display unit 205 to display, for example, the composite screen 21 shown in FIG.

[0191] The composite synthesis process generated by the synthesis unit 204 is displayed in a graph form on the composite screen 21. In other words, a composite synthesis flow is displayed. For example, in the composite synthesis flow shown in FIG. 16, the synthesis conditions "5h", "1223K", and "300MPa" included in the synthesis process of paper ID "0001" are replaced with the comprehensive synthesis conditions "4-7h", "900-1800K", and "250-320MPa", respectively. Furthermore, in the composite synthesis flow shown in FIG. 16, the characteristic value "1.30×10 -4 S / cm" is the comprehensive characteristic value of "1.30×10 -4 -1.35×10 -4 It has been replaced with "S / cm".

[0192] In addition, in the compound synthesis flow shown in FIG. 16, the comprehensive synthesis condition and the comprehensive characteristic value are displayed in a different manner from other synthesis conditions, processing words, and material words. For example, the color, thickness, size, or shape of the frame surrounding each of the comprehensive synthesis condition and the comprehensive characteristic value may be different from the frame surrounding other synthesis conditions. Alternatively, the color, size, or shape of the display area of ​​each of the comprehensive synthesis condition and the comprehensive characteristic value may be different from the display area of ​​other synthesis conditions. This makes it possible to present the comprehensive synthesis condition and the comprehensive characteristic value to the user in an easy-to-understand manner by distinguishing them from other synthesis conditions.

[0193] In this way, in this embodiment, the synthesis conditions used in each document are displayed as a comprehensive synthesis condition, so that by looking at the comprehensive synthesis condition, it is possible to easily search for a synthesis process for a new material. In addition, in this embodiment, even if the characteristic values ​​of the target material differ due to the different synthesis conditions used in each document, these characteristic values ​​are displayed as a comprehensive characteristic value, so that by looking at the comprehensive characteristic value, it is possible to easily search for a synthesis process for a new material.

[0194] [Processing Operation] FIG. 17A is a flowchart showing an example of the processing operation of composite process generation device 100 in this embodiment.

[0195] First, the document selection unit 101 selects one document as an extraction target document from a plurality of documents stored in the document storage unit 102 (step S11).

[0196] Next, the ingredient word extraction unit 103 extracts ingredient words from the extraction target document using the ingredient word dictionary shown in (a) of FIG. 3A (step S12). The process word extraction unit 104 extracts process words from the extraction target document using the process word dictionary shown in (b) of FIG. 3A (step S13). The condition extraction unit 105 extracts synthesis conditions from the extraction target document using the condition word dictionary shown in (a) of FIG. 3B (step S14). The characteristic value extraction unit 106 extracts characteristic values ​​of the target ingredient from the extraction target document using the characteristic value dictionary shown in (b) of FIG. 3B (step S15). Note that the processes in steps S12 to S15 may be performed in any order, or may be executed in parallel.

[0197] Next, the generating unit 107 generates a synthesis process by associating the extracted material words, processing words, synthesis conditions, and characteristic values ​​with each other (step S16). Furthermore, the generating unit 107 stores the generated synthesis process in the synthesis process storage unit 201 (step S17).

[0198] Next, document selection unit 101 determines whether all documents stored in document storage unit 102 have been selected as documents to be extracted (step S18). If it is determined that all documents have not been selected (No in step S18), document selection unit 101 repeats the process of step S11. That is, document selection unit 101 selects, as documents to be extracted, documents that have not yet been selected from among the multiple documents stored in document storage unit 102. On the other hand, if document selection unit 101 determines that all documents have been selected (No in step S18), combined process generation device 100 ends the combined process generation process.

[0199] FIG. 17B is a flowchart showing an example of the processing operation of synthesis process analysis apparatus 200 in this embodiment.

[0200] First, search condition input unit 202 accepts search conditions according to a signal indicating a user's operation result output from operation device 340 (step S21). Then, search unit 203 searches for a composite process matching the search conditions from among a plurality of composite processes stored in composite process storage unit 201 (step S22). Then, search unit 203 determines whether or not a composite process matching the search conditions is found by the search in step S22 (step S23). Here, if search unit 203 determines that a composite process matching the search conditions is not found (No in step S23), composite process analysis device 200 ends the analysis process of the composite process. On the other hand, if search unit 203 determines that a composite process matching the search conditions is found (Yes in step S23), it further determines whether or not there are a plurality of composite processes found (step S24). Here, if searching unit 203 determines that the number of found composite processes is one and not multiple (No in step S24), it causes display unit 205 to display the composite process as a composite flow in response to a user operation (step S27). On the other hand, if searching unit 203 determines that the number of found composite processes is multiple (Yes in step S24), compounding unit 204 compounds the multiple found composite processes in response to a user operation (step S25). Then, compounding unit 204 causes display unit 205 to display the composite composite process generated by the compounding as a composite composite flow (step S26).

[0201] As described above, in this embodiment, a synthesis process, which is a processing procedure for synthesizing a target material from a plurality of starting materials, can be easily generated from a document. In addition, since a composite synthesis process is generated from the generated plurality of synthesis processes, it is possible to display common parts and different parts of many synthesis processes in an easily understandable manner. This makes it possible to appropriately support the search for a synthesis process for a new material. In addition, when a plurality of synthesis processes generated from the latest documents are combined, knowledge regarding material development can be kept up to date. In addition, by utilizing a huge number of documents, it is possible to support the search for a synthesis process that covers a wide range of materials.

[0202] In this embodiment, the multiple combined processes generated by the generation unit 107 and the composite combined process generated by the combination unit 204 are each configured to be graphable. Then, the display unit 205 displays the composite combined process generated by the combination unit 204 as a graph. That is, in this embodiment, the multiple process information generated by the generation unit 107 and the composite process information generated by the combination unit 204 are each configured to be graphable. Then, the display unit 205, which is an example of an output unit, displays the composite process information generated by the combination unit 204 as a graph.

[0203] In this way, since a plurality of composite processes are configured as graphable structured data, a composite composite process can also be easily generated as graphable structured data. Furthermore, since the composite composite process is displayed as a graph such as a flow chart, the user can appropriately understand the composite composite process.

[0204] (Variation 1) In the first embodiment, the material words of the starting material, the processing words, and the synthesis conditions are shown in the same manner in the compound synthesis flow. In this modification, the material words of the starting material, the processing words, and the synthesis conditions are shown in different manners.

[0205] FIG. 18 is a diagram showing an example of a composite screen displayed on display unit 205 according to this modification.

[0206] When the composite unit 204 according to this modification generates a composite composite process as described above, it displays the composite screen 22 shown in FIG. 18 on the display unit 205 instead of the composite screen 21 shown in FIG.

[0207] In this composite screen 22, the composite synthesis process generated by the synthesis unit 204 is displayed as a composite synthesis flow. In the composite synthesis flow according to this modification, the material words of the starting material, the material words of the target material, the processing words, the synthesis conditions, and the characteristic values ​​are displayed in different modes. For example, as shown in FIG. 18, the material words of the starting material are surrounded by a solid frame, the processing words are surrounded by a double frame of solid and dashed lines, the synthesis conditions are surrounded by a dashed frame, the material words of the target material are surrounded by a thick solid frame, and the characteristic values ​​are surrounded by a double frame of solid lines. In this way, the material words of the starting material, the material words of the target material, the processing words, the synthesis conditions, and the characteristic values ​​may be surrounded by frames different from each other. Alternatively, the color, size, or shape of the display areas of the material words of the starting material, the material words of the target material, the processing words, the synthesis conditions, and the characteristic values ​​may be made different from each other.

[0208] This makes it possible to distinguish between the material words of the starting material, the material words of the target material, the processing words, the synthesis conditions, and the characteristic values ​​and present them to the user in an easily understandable manner.

[0209] (Variation 2) In the first embodiment, a comprehensive synthesis condition is generated by comprehensively including a plurality of synthesis conditions that are different from one another. In this modification, the plurality of synthesis conditions are not comprehensively included, but are partially included by clustering.

[0210] FIG. 19 is a diagram for explaining the process of the composite unit 204 according to this modification.

[0211] In the present modified example, when a plurality of different synthesis conditions each indicate a plurality of different numerical values, the compositing unit 204 generates a plurality of subsets by performing clustering on a set including the plurality of different numerical values, and generates a comprehensive synthesis condition from each of the plurality of subsets. In other words, in the present modified example, when a plurality of different synthesis conditions 1 to n each indicate a plurality of different numerical values ​​1 to n, the compositing unit 204 generates a plurality of subsets by performing clustering on a set including the plurality of different numerical values ​​1 to n, and generates a comprehensive synthesis condition from each of the plurality of subsets.

[0212] For example, as shown in FIG. 19, the search unit 203 searches for six synthesis processes. Here, the synthesis conditions of pressure associated with the processing word "press" are different from one another in the six synthesis processes, such as "300MPa", "250MPa", "320MPa", "450MPa" and "460MPa". At this time, the composite unit 204 generates two subsets by performing clustering on the set including the different pressure values. For the clustering, an existing method (for example, the k-means method) is used. As a result, a first subset including "300MPa", "250MPa" and "320MPa" and a second subset including "450MPa" and "460MPa" are generated. In FIG. 19, the first subset is indicated by a dashed line frame, and the second subset is indicated by a dashed line frame. The composite unit 204 generates a comprehensive synthetic pressure condition "250-320 MPa" from the first subset, and generates a comprehensive synthetic pressure condition "450-460 MPa" from the second subset.

[0213] The composite unit 204 also derives a comprehensive characteristic value “1.30×10 -4 -1.35×10 -4 S / cm" and a comprehensive property value of "1.41×10 -4 -1.42×10 -4 S / cm.

[0214] Then, the composite unit 204 selects, for example, a composite process with a paper ID of "0001" as a composite process to be replaced, replaces the composite conditions included in the composite process to be replaced with two comprehensive composite conditions, and replaces the characteristic values ​​included in the composite process to be replaced with two comprehensive characteristic values. This generates a composite composite process that indicates the parts common to the six composite processes and also indicates multiple parts different from each other in the six composite processes.

[0215] FIG. 20A is a diagram showing an example of a composite screen displayed on display unit 205 according to this modification.

[0216] When the composite unit 204 generates the composite combined process as described above, it displays, for example, a composite screen 23a shown in Fig. 20A on the display unit 205. On the composite screen 23a, the composite combined process generated by the composite unit 204 is displayed as a composite combined flow.

[0217] For example, in the composite synthesis flow shown in FIG. 20A, the pressure synthesis condition "300 MPa" included in the synthesis process of paper ID "0001" is replaced with the comprehensive synthesis conditions "250-320 MPa" and "450-460 MPa". Furthermore, in the composite synthesis flow shown in FIG. 20A, the characteristic value "1.30×10 -4 S / cm" is the comprehensive characteristic value of "1.30×10 -4 -1.35×10 -4 S / cm" and "1.41×10 -4 -1.42×10 -4 It has been replaced with "S / cm".

[0218] For example, when the variance of multiple different synthesis conditions is small, a comprehensive synthesis condition may be generated by comprehensively including all of the synthesis conditions as in the above-mentioned embodiment 1. However, when the variance is large, it is possible to easily find an unused synthesis condition by generating multiple comprehensive synthesis conditions by performing clustering as in this modified example.

[0219] FIG. 20B is a diagram showing another example of a composite screen displayed on display unit 205 according to this modification.

[0220] When the composite unit 204 generates the composite combined process as described above, the composite unit 204 may display, for example, a composite screen 23b shown in Fig. 20B on the display unit 205. The composite screen 23b displays the composite combined process generated by the composite unit 204 as a composite combined flow.

[0221] For example, the compound synthesis flow shown in FIG. 20B branches into two from the processing word "dry." In other words, two branch flows are connected to the processing of the processing word "dry." The two branch flows are connected to the processing words "press" and "sinter" and the material word "Li" of the target material, respectively. 1.1 Mg 0.6 Ti 1.9 The processing word "press" included in the first branch flow of the two branch flows is associated with the comprehensive synthesis condition "250-320MPa", and the processing word "press" included in the second branch flow is associated with the comprehensive synthesis condition "450-460MPa". Furthermore, the material word "Li 1.1 Mg 0.6 Ti 1.9 (PO4)3" has a global characteristic value of 1.30×10 -4 -1.35×10 -4 The material word "Li S / cm" of the target material in the second branch flow is associated with 1.1 Mg 0.6 Ti 1.9 (PO4)3" has a global characteristic value of 1.41×10 -4 -1.42×10 -4 "S / cm" is associated with it.

[0222] This makes it possible to confirm that there are two types of synthesis processes in the multiple documents found, and to confirm the characteristic values ​​of the target materials for each synthesis process.

[0223] In this manner, in this modified example, a plurality of comprehensive synthesis conditions are generated by clustering, so that the tendency of synthesis conditions used in each document can be grasped in more detail.

[0224] (Variation 3) In the first embodiment, a composite synthesis process is generated by combining synthesis processes whose processing words are all common. In this modification, a composite synthesis process is generated by combining synthesis processes whose processing words are different from each other.

[0225] Each of Fig. 21A and Fig. 21B is a diagram showing an example of a synthesis process to be combined in this modified example. For example, when the synthesis process search screen 2a shown in Fig. 12 is displayed by the display unit 205, the user selects the search start button 2e without inputting the name of the process (i.e., the process word) in the process input field. That is, the material words "Li2CO3", "MgO", "TiO2" and "NH4H2PO4" of the starting materials and the material word "Li 1.1 Mg 0.6 Ti 1.9 The search for synthesis processes is started using "(PO4)3", the publication period "2010 / 1 / 1" to "2019 / 12 / 31", and the number of citations "10" or more as search criteria.

[0226] By this search, for example, a synthesis process of paper ID "0001" shown in FIG. 21A and a synthesis process of paper ID "0142" shown in FIG. 21B are found.

[0227] In these synthesis processes, the starting material and the target material are the same. However, the processes are different from each other. That is, the synthesis process of paper ID "0001" includes the processed words "mix", "heat", "mill", "dry", "press" and "sinter", but the synthesis process of paper ID "0142" includes the processed words "mix", "heat", "mill" and "heat". In other words, the synthesis process of paper ID "0001" and the synthesis process of paper ID "0142" include the common processed words "mix", "heat" and "mill", but also include different processed words from each other. The processed words "dry", "press" and "sinter" included in the synthesis process of paper ID "0001" are different from the processed word "heat" included in the synthesis process of paper ID "0142".

[0228] In addition, the synthesis process of the paper ID "0001" and the synthesis process of the paper ID "0142" include a common processing word "heat". However, the synthesis conditions associated with the common processing word "heat" are different in these two synthesis processes. That is, in the synthesis process of the paper ID "0001", the synthesis conditions are "5h" and "1223K", and in the synthesis process of the paper ID "0142", the synthesis conditions are "7h" and "1200K". Therefore, the combining unit 204 merges these different synthesis conditions.

[0229] FIG. 22 is a diagram for explaining the process of the composite unit 204. As shown in FIG.

[0230] The combining unit 204 merges different combining conditions in each of the two combining processes. For example, the combining unit 204 generates a comprehensive time combining condition "5-7h" by merging the time combining conditions "5h" and "7h" associated with the common processing word "heat" in each of the two combining processes. Similarly, the combining unit 204 generates a comprehensive temperature combining condition "1200-1223K" by merging the temperature combining conditions associated with the common processing word "heat".

[0231] The compound unit 204 according to this modification generates a compound synthesis process that includes each of the comprehensive synthesis conditions described above and that includes processing words that are different from each other.

[0232] FIG. 23 is a diagram showing an example of a composite screen displayed on display unit 205 according to this modification.

[0233] When the composite unit 204 generates the composite combined process as described above, it displays, for example, a composite screen 24 shown in Fig. 23 on the display unit 205. On the composite screen 24, the composite combined process generated by the composite unit 204 is displayed as a composite combined flow.

[0234] For example, in the composite synthesis flow shown in FIG. 23, the synthesis conditions "5h" and "1223K" included in the synthesis process of the paper ID "0001" are replaced with the comprehensive synthesis conditions "5-7h" and "1200-1223K", respectively.

[0235] Furthermore, in the composite synthesis flow shown in Fig. 23, a part of the synthesis process of paper ID "0142" is added to the synthesis process of paper ID "0001". Specifically, the composite synthesis flow shown in Fig. 23 branches into two from the processing word "mill" that is common to the two synthesis processes described above. In other words, a branch flow that is part of the synthesis process of paper ID "0142" is connected after the processing word "mill". This branch flow is a flow in which the target material is generated through processing the processing word "heat".

[0236] In this modified example, the composite unit 204 generates a composite composite process by comprehensively associating each of the different parts with the common part in two composite processes so that the process flow branches from the common part to each of the different parts. That is, in this modified example, in (a-2), the process word extraction unit 104 extracts at least one process word i including the above-mentioned process word i from the document i, and in (a-3), the condition extraction unit 105 extracts at least one synthesis condition i including the above-mentioned synthesis condition i from the document i. Then, the composite unit 204 generates composite process information by associating each of the different parts with the common part so that the process flow of material production branches from the above-mentioned common part to each of the above-mentioned different parts.

[0237] As a result, in this modified example, for example, when a common processing word and a plurality of different processing words are included in a plurality of synthesis processes, a composite synthesis process is generated so that the flow of material generation branches from the processing of the common processing word to the processing of the plurality of different processing words. As a result, even if a plurality of synthesis processes include a plurality of different processing words, the common processing and the plurality of different processing can be displayed in an easy-to-understand manner. In addition, the plurality of different processing can be displayed in an easy-to-comprehend manner.

[0238] In the examples shown in Figures 21A to 23, each of the multiple pieces of composite process information used to generate the composite composite process information includes multiple processing words and multiple synthesis conditions. However, each of the multiple pieces of composite process information may include one processing word and one synthesis condition. In other words, each of the multiple pieces of composite process information may be generated by extracting one processing word and one synthesis condition from one document.

[0239] For example, as shown in Fig. 21A, multiple starting material words, multiple processing words, multiple synthesis conditions, target material words, and characteristic values ​​are extracted from a document with a paper ID of "0001", and synthesis process information is generated using these. However, instead of multiple processing words and multiple synthesis conditions, for example, one processing word "dry" and one synthesis condition "393K" corresponding to that "dry" may be extracted.

[0240] Similarly, as shown in Fig. 21B, multiple starting material words, multiple processing words, multiple synthesis conditions, target material words, and characteristic values ​​are extracted from the document with the paper ID "0142", and synthesis process information is generated using them. However, for example, instead of multiple processing words and multiple synthesis conditions, one processing word "heat" and one synthesis condition "450K" corresponding to that "heat" may be extracted.

[0241] In this case, the synthesis process information of the paper ID "0001" and the synthesis process information of the paper ID "0142" have multiple starting material words in common. On the other hand, between these synthesis process information, the processing word "dry" and the synthesis condition "393K" are different from the processing word "heat" and the synthesis condition "450K". Therefore, in the composite process information generated from these synthesis process information, there are flows branching from the common part (i.e., multiple starting material words) to the processing word "dry" and the synthesis condition "393K" and flows branching to the processing word "heat" and the synthesis condition "450K". In other words, even when one processing word and one synthesis condition are extracted from each of multiple documents, the composite process information can be generated by associating each of the different parts with the common part so that the common part branches to the above-mentioned different parts.

[0242] (Variation 4) In the first embodiment, the name of the organization to which the author belongs, that is, the institution, is not used to search for synthesis processes. In this modification, the name of the organization or the institution is used to search for synthesis processes.

[0243] FIG. 24 is a diagram showing an example of a composite process search screen displayed by display unit 205 according to this modification.

[0244] The search unit 203 according to this modification causes the display unit 205 to display a combination process search screen 2f shown in FIG. 24 in response to a signal indicating a result of a user's operation output from the operation device 340, for example.

[0245] The synthesis process search screen 2f differs from the synthesis process search screen 2a in the first embodiment in that it further includes an organization name as a search condition. That is, the synthesis process search screen 2f includes a search condition window 2g, which includes an input field for inputting an organization name. The organization name is the name of the organization (i.e., the organization name) to which the author who issued or published the document used in generating the synthesis process belongs. In this modification, the bibliographic information added to each synthesis process stored in the synthesis process storage unit 201 includes the organization name.

[0246] When such a synthesis process search screen 2f is displayed, the user performs input processing for each input field by operating the operation device 340. For example, as shown in FIG. 24, the user inputs "University A" into the input field for the institution name. Furthermore, the user inputs the material words "Li2CO3", "MgO", "TiO2" and "NH4H2PO4" into the four input fields for the starting materials, respectively. Furthermore, the user inputs the material word "Li 1.1 Mg 0.6 Ti 1.9 Enter "(PO4)3" in the "Publication Period" field, enter "2010 / 1 / 1" to "2019 / 12 / 31" in the "Publication Period" field, and enter "10" or more in the "Number of Citations" field.

[0247] Then, the user operates the operation device 340 to select search start button 2e. As a result, the search condition input unit 202 accepts the name of the institution, the name of the starting material, the name of the target material, the publication period, and the number of citations input in the search condition window 2g as search conditions. Then, the search condition input unit 202 outputs the search conditions to the search unit 203. Upon acquiring the search conditions from the search condition input unit 202, the search unit 203 searches for a synthesis process that matches the search conditions from among the multiple synthesis processes stored in the synthesis process storage unit 201.

[0248] When the compounding unit 204 finds a plurality of compound processes by such a search, it generates a composite compound process by compounding the compound processes in the same manner as in the above-mentioned embodiment 1 or its modification. The composite compound process generated in this manner is a composite compound process for University A.

[0249] Here, for example, when "Company B" is entered instead of "Company A" in the input field for the organization name on the composite process search screen 2f shown in FIG.

[0250] Each of FIG. 25A and FIG. 25B is a diagram showing an example of a combined screen displayed on display unit 205 according to this modification.

[0251] When a composite composite process for University A is generated, the composite unit 204 causes the display unit 205 to display, for example, a composite screen 21a for University A shown in Fig. 25A. The composite composite process for University A generated by the composite unit 204 is displayed as a composite composite flow on the composite screen 21a. On the other hand, when a composite composite process for Company B is generated, the composite unit 204 causes the display unit 205 to display, for example, a composite screen 21b for Company B shown in Fig. 25B. The composite screen 21b displays the composite composite process for Company B generated by the composite unit 204 as a composite composite flow.

[0252] The synthesis process of the target material varies depending on the institution or organization, such as a university or a company, that performs the experiment, and the tendency to perform the experiment may also differ depending on the facilities or development strategies of the institutions. Therefore, in this modified example, as shown in Figs. 25A and 25B, a composite synthesis process is displayed for each institution (e.g., a university or a company) to which the author of a document such as a paper belongs. This allows the user to compare and understand the composite synthesis process for each institution. In the example shown in Fig. 25A, the synthesis condition associated with the processing word "heat" in the composite synthesis process for University A is a short time and a low temperature. On the other hand, in the example shown in Fig. 25B, the synthesis condition associated with the processing word "heat" in the composite synthesis process for Company B is a long time and a high temperature. As a result, the characteristic value of the target material in the composite synthesis process for Company B tends to be larger than the characteristic value of the target material in the composite synthesis process for University A. In this manner, in this modified example, the composite synthesis process for each institution can be compared.

[0253] In this modified example, similarly to the above embodiment, multiple synthesis processes are searched and synthesized using the publication period as a search condition. Therefore, it is possible to classify the synthesis processes by the year in which the document such as a paper was published, and it is also possible to confirm how the synthesis conditions change by year.

[0254] (Variation 5) In this modification, an intermediate material is synthesized from a plurality of starting materials, and composite process information is generated that indicates a procedure for synthesizing a target material from the intermediate material.

[0255] FIG. 26 is a diagram showing an example of composite process information including synthesis of intermediate materials.

[0256] For example, as shown in FIG. 26, a generation device that is a synthesis process generation device 100 generates first process information from a document and generates second process information from another document. The first process information indicates a plurality of starting material words, a processing word, a synthesis condition, and a target material word in association with each other. Similarly to the first process information, the second process information also indicates a plurality of starting material words, a processing word, a synthesis condition, and a target material word in association with each other. Here, there are cases where the target material word of the first process information and one of the plurality of starting material words of the second process information are the same word and have something in common.

[0257] In such a case, the composite unit 204 in this modification may generate composite process information indicating a procedure for generating a target material indicated by the second process information from a plurality of starting materials indicated by the first process information. In other words, the target material indicated by the first process information is treated as an intermediate material for generating the target material indicated by the second process information.

[0258] In this modified example, among the multiple process information, the target material word of the first process information and the starting material word of the second process information are the same material word and are a common part of the multiple process information. In such a case, the composite process information generated by the composite unit 204 shows a procedure for synthesizing the materials indicated by the same material word, which is the common part, as intermediate materials from the multiple starting materials indicated by the first process information, and further generating the target material indicated by the second process information from the intermediate material.

[0259] This makes it possible to generate composite process information that indicates a procedure not described in any document. Therefore, when such composite process information is displayed, it is possible to further support the search for a new composite process.

[0260] (Embodiment 2) The synthesis process analysis apparatus in the second embodiment accepts synthesis conditions from a user, and estimates characteristic values ​​of a target material according to the accepted synthesis conditions.

[0261] 27 is a block diagram showing an example of a functional configuration of a synthesis process search support system according to the present embodiment. Note that, among the components in the present embodiment, the same components as those in the first embodiment are given the same reference numerals as those in the first embodiment, and detailed description thereof will be omitted.

[0262] The composite process search support system 11 in this embodiment includes a composite process generation device 100 and a composite process analysis device 240.

[0263] Synthetic process generation device 100 generates a synthetic process in the same manner as in embodiment 1. Moreover, condition extraction unit 105 and characteristic value extraction unit 106 of synthetic process generation device 100 in this embodiment output the extracted synthesis conditions and characteristic values ​​to synthetic process analysis device 240, respectively.

[0264] Synthesis process analysis device 240 in this embodiment includes the components of synthesis process analysis device 200 in the first embodiment, and further includes association accumulation unit 206, synthesis condition input unit 207, and characteristic value estimation unit 208.

[0265] [Related Storage Section] The association storage unit 206 is a recording medium having a storage capacity for storing the synthesis conditions extracted by the condition extraction unit 105 and the characteristic values ​​extracted by the characteristic value extraction unit 106 in association with each other. In this association storage unit 206, at least one synthesis condition and a characteristic value are associated with each document or each synthesis process generated from the document. The association storage unit 206 may be, for example, a hard disk, a RAM, a ROM, or a semiconductor memory. The association storage unit 206 may be volatile or non-volatile.

[0266] [Composite condition input section] The synthesis condition input unit 207 receives, as an input synthesis condition, a synthesis condition corresponding to a signal indicating a user's operation result output from the operation device 340. That is, the synthesis condition input unit 207 receives an input synthesis condition that is a processing condition.

[0267] [Characteristic value estimation section] When a composite synthesis process is generated and input synthesis conditions are received by the synthesis condition input unit 207, the characteristic value estimation unit 208 estimates a characteristic value of a target material according to the input synthesis conditions. The target material is a common target material in the multiple synthesis processes used to generate the composite synthesis process. Therefore, the characteristic value estimation unit 208 in this embodiment estimates a characteristic value according to the input synthesis conditions of the common target material. In other words, the characteristic value estimation unit 208 estimates a characteristic value according to the input synthesis conditions of the target material of the common target material word.

[0268] Specifically, the characteristic value estimation unit 208 acquires, from the association accumulation unit 206, the synthesis conditions and characteristic values ​​included in each of the synthesis processes used in generating the composite synthesis process. Then, the characteristic value estimation unit 208 estimates the characteristic value of the target material according to the input synthesis conditions accepted by the synthesis condition input unit 207 using the acquired synthesis conditions and characteristic values. This estimation of the characteristic value may use a multiple regression analysis method or the like. For example, the characteristic value estimation unit 208 models a formula that outputs a characteristic value for the input synthesis conditions from the synthesis conditions and characteristic values ​​acquired from the association accumulation unit 206 according to the multiple regression analysis method. Then, the characteristic value estimation unit 208 derives the characteristic value by substituting the input synthesis conditions accepted by the synthesis condition input unit 207 into the modeled formula. This estimates the characteristic value. Note that instead of the multiple regression analysis method, simultaneous equations may be used, and the characteristic value may be estimated by an interpolation process. Note that the interpolation may be an interpolation or an extrapolation.

[0269] In this manner, the characteristic value estimation unit 208 in this embodiment estimates a characteristic value according to the input synthesis condition based on the relationship between the synthesis condition included in each of the multiple synthesis processes and the characteristic value included in each of the multiple synthesis processes. In other words, the characteristic value estimation unit 208 in this embodiment estimates a characteristic value according to the input synthesis condition based on the relationship between the synthesis condition i included in each of the multiple process information and the characteristic value i included in each of the multiple process information.

[0270] FIG. 28 is a diagram showing an example of a composite screen displayed on display unit 205 in this embodiment.

[0271] Similar to the first embodiment, the composite unit 204 in this embodiment generates a composite composite process from the four composite processes shown in Fig. 14A to Fig. 14D, and then displays, for example, a composite screen 25 shown in Fig. 28 on the display unit 205. On the composite screen 25, the composite composite process generated by the composite unit 204 is displayed as a composite composite flow.

[0272] In the composite synthesis flow of this embodiment, the comprehensive synthesis condition is shown as an input range. Furthermore, the comprehensive characteristic value is not shown. For example, the comprehensive synthesis condition "4-7h" is shown as the input range of the synthesis condition for time in processing the processing word "heat", and the comprehensive synthesis condition "900-1800K" is shown as the input range of the synthesis condition for temperature in the processing. Similarly, the comprehensive synthesis condition "250-320K" is shown as the input range of the synthesis condition for pressure in processing the processing word "press".

[0273] Here, the composite screen 25 is provided with input fields 25a to 25c (framed with thick solid lines in FIG. 28) for inputting the synthesis conditions for each input range. The user inputs the synthesis conditions into the input fields 25a to 25c by operating the operation device 340. The synthesis conditions input in this way are accepted as input synthesis conditions by the synthesis condition input unit 207. Then, when the input of the synthesis conditions into all the input fields 25a to 25c is completed, the characteristic value estimation unit 208 estimates the characteristic value of the target material according to the synthesis conditions input into each of the input fields 25a to 25c. Then, the characteristic value estimation unit 208 causes the composite screen 25 of the display unit 205 to display the estimated characteristic value. The characteristic value estimated in this way is displayed on the composite screen 25 of the display unit 205 as the material word "Li" of the target material on the composite screen 25. 1.1 Mg 0.6 Ti 1.9 28, the estimated characteristic values ​​are placed in the thick dashed frame. In this manner, display unit 205 displays the characteristic values ​​estimated by characteristic value estimation unit 208. Display unit 205, which is an example of an output unit in this embodiment, displays the characteristic values ​​as described above, but the output unit may also output the characteristic values ​​estimated by characteristic value estimation unit 208.

[0274] For example, the synthesis condition input unit 207 receives, as input synthesis conditions, the synthesis condition of time "5h" and the synthesis condition of temperature "1300K" for the processing of the processing word "heat", and the synthesis condition of pressure "280MPa" for the processing of the processing word "press". In this case, the characteristic value estimation unit 208 estimates the characteristic value of the target material "1.32×10 -4 S / cm" is estimated and displayed on the display unit 205.

[0275] FIG. 29 is a diagram for explaining a method of inputting data into the input field of the combined screen 25. As shown in FIG.

[0276] For example, when inputting a time synthesis condition in the input field 25a, the user may input an arbitrary synthesis condition by operating the operation device 340, or may select one of multiple candidates as the synthesis condition.

[0277] Specifically, when the composite unit 204 generates a comprehensive composite condition, the composite screen 25 is provided with an input field 25a in which at least one numerical value within the range indicated by the comprehensive composite condition can be selected as a candidate for the composite condition. For example, the input field 25a is configured by a pull-down menu. When the comprehensive composite condition of time is "4-7h", the input range is "4-7h". Therefore, when the input field 25a is selected by the user, 4h, 5h, 6h, and 7h are displayed as candidates for the composite condition of time. The user selects one of the candidates by operating the operation device 340. As a result, the selected candidate is input as a composite condition in the input field 25a. The composite condition thus input is accepted by the composite condition input unit 207 as an input composite condition.

[0278] In this embodiment, when the user sets the input synthesis conditions, the characteristic values ​​of the target material are estimated and displayed, so that, for example, a synthesis process of a material having a new characteristic value can be easily searched for. Therefore, the user can grasp the characteristic values ​​of the target material for any synthesis conditions without conducting an experiment.

[0279] (Embodiment 3) The synthesis process analysis apparatus in the third embodiment presents information on facilities capable of realizing the comprehensive synthesis conditions.

[0280] 30 is a block diagram showing an example of a functional configuration of a synthesis process search support system according to the present embodiment. Note that, among the components in the present embodiment, the same components as those in the first embodiment are given the same reference numerals as those in the first embodiment, and detailed description thereof will be omitted.

[0281] The composite process search support system 12 in this embodiment includes a composite process generation device 100 and a composite process analysis device 250.

[0282] A composite process analysis apparatus 250 in this embodiment includes the components of composite process analysis apparatus 200 in the first embodiment, and further includes an equipment characteristic storage unit 211 and an equipment search unit 212.

[0283] [Facility Characteristics Storage Department] The equipment characteristic storage unit 211 is a recording medium that holds equipment information. The equipment information indicates, for each of a plurality of pieces of equipment, a list of processing conditions that can be realized by the equipment. The equipment characteristic storage unit 211 may be, for example, a hard disk, a RAM, a ROM, or a semiconductor memory. The equipment characteristic storage unit 211 may be volatile or non-volatile.

[0284] [Facility Search Section] The facility search unit 212 searches for facilities that can perform processing under the comprehensive synthesis conditions generated by the composition unit 204 by referring to the above-mentioned facility information.

[0285] Moreover, the display unit 205 in this embodiment displays information about the facility searched for by the facility search unit 212. Note that the display unit 205, which is an example of an output unit in this embodiment, displays information about the facility as described above, but the output unit may output information about the facility searched for by the facility search unit 212.

[0286] FIG. 31 is a diagram showing an example of facility information in the present embodiment.

[0287] The equipment information indicates the performance value and price of each piece of equipment, as shown in FIG. 31, for example. Equipment is identified by the manufacturer name and model number. The performance value is a feasible processing condition. For example, the equipment information indicates that equipment with the manufacturer name "Company A" and model number "AS-0122" has a performance value of "260-300 MPa" as a feasible condition for pressurization processing, and the price of the equipment is "500,000 yen."

[0288] FIG. 32 is a diagram showing an example of a composite screen displayed on display unit 205 in this embodiment.

[0289] Similar to the first embodiment, the composite unit 204 in this embodiment generates a composite composite process from the four composite processes shown in Fig. 14A to Fig. 14D, and then displays the composite screen 26 shown in Fig. 32 on the display unit 205. The composite composite process generated by the composite unit 204 is displayed as a composite composite flow on the composite screen 26.

[0290] Furthermore, in this embodiment, the equipment search unit 212 displays information on the searched equipment on the composite screen 26. For example, the composite unit 204 generates a comprehensive composite pressure condition of "250-320 MPa". At this time, the equipment search unit 212 searches for equipment capable of performing pressurization processing with the comprehensive composite pressure condition of "250-320 MPa" by, for example, referring to the equipment information shown in FIG. 31. As a result, the equipment search unit 212 finds equipment with a manufacturer name of "Company B" and a model number of "YP-0134". Then, the equipment search unit 212 displays the manufacturer name of "Company B" and the model number of "YP-0134" and, for example, a price of "800,000 yen" associated with the equipment in the equipment information on the display unit 205 in association with the comprehensive composite condition of "250-320 MPa". It should be noted that the price does not have to be displayed.

[0291] In this embodiment, the composite synthesis flow is displayed, and the user can easily be informed of the facilities that can perform processing under the comprehensive synthesis conditions included in the composite synthesis flow. As a result, the user can easily arrange the facilities required for material synthesis.

[0292] In other words, in this embodiment, when the comprehensive synthesis conditions included in a composite synthesis flow and the comprehensive characteristic values ​​of the target material are displayed, it is possible to confirm what equipment can be used to verify the composite synthesis flow.

[0293] In this embodiment, information about the equipment is displayed simultaneously with the composite synthesis flow, but it does not have to be displayed simultaneously. For example, when a displayed comprehensive synthesis condition is selected by the user, information about the equipment corresponding to the comprehensive synthesis condition may be displayed. That is, when the equipment search unit 212 receives a signal indicating that a comprehensive synthesis condition included in the composite synthesis flow has been selected from the operation device 340, it searches for equipment that can perform processing under the comprehensive synthesis condition by referring to the equipment information. Then, the equipment search unit 212 displays information about the equipment on the display unit 205 in association with the comprehensive synthesis condition.

[0294] (Embodiment 4) The support device in this embodiment is a system that extracts and analyzes information from documents such as papers, thereby making the sales activities of a sales company that sells materials or equipment more efficient.

[0295] In the sales activities of a sales company, it is important to collect information on customers in advance and find customers who are likely to purchase the material or device in question in order to increase the success rate of a contract.

[0296] However, the main customers of the distributors, such as universities and companies, rarely disclose information about the materials or equipment they use in their experiments, making it difficult for the distributors to collect information about their customers.

[0297] Furthermore, in order to conduct sales activities efficiently, it is necessary to continually update information about customers. However, since sales activities at sales companies are mainly conducted manually, updating information about customers is costly.

[0298] Therefore, the support device in this embodiment can improve the efficiency of sales activities of sales companies by analyzing information extracted from paper data published by researchers belonging to an organization.

[0299] [System Overview] FIG. 33 is a diagram illustrating an example of a schematic configuration of the support device.

[0300] The support device 13 includes a first analysis device 110, a second analysis device 120, a thesis database 130, and a database group 140, all of which are connected to each other via a communication network 150.

[0301] The paper database 130 is also called a paper DB and is a recording medium that holds a plurality of paper data, each of which is a document. The paper database 130 corresponds to the document storage unit 102 in the first to third embodiments.

[0302] The database group 140 includes a plurality of databases, each of which is a recording medium used by at least one of the first analysis device 110 and the second analysis device 120. Each of these plurality of databases is configured as a recording medium, similar to the paper database 130. These recording media are, for example, hard disks. Note that the recording media may be recording media other than hard disks, such as RAM (Random Access Memory), ROM (Read Only Memory), or semiconductor memory. Furthermore, these recording media may be either volatile or non-volatile.

[0303] First analysis device 110 reads out a plurality of pieces of paper data from paper database 130 via communication network 150. Then, first analysis device 110 generates a data set corresponding to each of the plurality of pieces of paper data by extracting each piece of information from the paper data as customer information. First analysis device 110 stores the generated data set in database group 140 via communication network 150. Note that in this embodiment, the organization to which the author of the paper belongs is, for example, a university, research institute, or company, and is a customer of a sales company that sells products such as equipment or materials.

[0304] Second analysis device 120 reads the plurality of data sets from database group 140 via communication network 150, and estimates various information related to the purchase of the above-mentioned product by the customer from the plurality of data sets. In this embodiment, second analysis device 120 is used, for example, by the above-mentioned sales company or a person belonging to the sales company.

[0305] The first analysis device 110, the second analysis device 120, the paper database 130, and the database group 140 may be installed in the same facility, or may be located in multiple facilities that are separate from one another. The first analysis device 110 and the second analysis device 120 may be used by different users, or may be used by the same user.

[0306] [Paper data] FIG. 34 is a diagram showing an example of the contents of the paper bib.

[0307] The paper data in this embodiment is composed of a combination of two files. One of the two files is a paper PDF (Portable Document Format), and the other is a paper bib. The paper PDF is a file that is composed of a PDF file, and the contents of the paper are described, for example, in a natural language. This paper PDF may also contain text data. The paper bib is a file with the extension "bib" and is also called a BIB file.

[0308] The paper bib is metadata of a paper, as shown in FIG. 34, for example. In such a paper bib, the title of the paper, the name of the author of the paper, the name of the organization to which the author belongs (Affiliation), the publication date of the paper, etc. are described. Hereinafter, the name of the author will be also called the author name, and the name of the organization will be also called the organization name, research organization name, or affiliated organization name. Furthermore, the paper bib is structured data.

[0309] Therefore, the first analysis device 110 in this embodiment can easily extract the above-mentioned author names, organization names, publication dates, and the like from the paper bib through processing by a computer program.

[0310] FIG. 35 is a diagram showing an example of the contents of a thesis PDF.

[0311] As shown in (a) of FIG. 35, a paper PDF contains, for example, information about the equipment used in the experiments of the paper and information about the equipment provider that sold or manufactured the equipment. The information about the equipment is also called equipment information, and is, for example, the name, alias, model, or model number of the equipment. The information about the equipment provider is, for example, the name of the equipment provider. The equipment provider is, specifically, the sales company or manufacturer of the equipment, and can also be said to be the purchaser of the equipment.

[0312] In the example shown in Fig. 35(a), the paper PDF describes the device model number "TN-200" and the name of the supplier of the device "Korea Chem Inc." Furthermore, the paper PDF describes the device model number "XRD, Gaku Mini II" and the name of the supplier of the device "Japan Chem Inc."

[0313] Furthermore, as shown in (b) of FIG. 35, a paper PDF contains, for example, information about the materials used in the experiments of the paper and information about the material provider that sold or manufactured the materials. The information about the materials is also called material information, and is, for example, the name or alias of the material. The information about the material provider is, for example, the name of the material provider. The material provider is, specifically, the sales company or manufacturer of the material, and can also be said to be the purchaser of the material.

[0314] In the example shown in (b) of Fig. 35, the paper PDF describes "China Industry" as the name of the supplier of the material named "Octylamine". Furthermore, the paper PDF describes "Sako Pure Chemicals Co., Ltd" as the name of the supplier of the material named "Acetone and methanol".

[0315] FIG. 36 is a diagram showing an example of another content of the thesis PDF.

[0316] The paper PDF further includes information about the project supporting the experiments in the paper, as shown in Fig. 36. The information about the project is also called project information, and is, for example, the name or identification number of the project.

[0317] In the example shown in Fig. 36, the paper PDF describes "the National Material Science Foundation of Japan" as the project name, and "No. 11112222 and 33334444" as the project identification number. Furthermore, the paper PDF describes "the the Specialized Research Fund for the Doctoral Program of Higher Education of Japan" as the project name, and "No. ABC111" as the project identification number.

[0318] [System Details] FIG. 37 is a diagram showing an example of a detailed configuration of the support device 13. As shown in FIG.

[0319] First analysis device 110 includes a composite process extraction unit 111 and an extraction unit 112 .

[0320] Like composite process generation device 100 in the above-mentioned first to third embodiments, composite process extraction unit 111 generates a composite process from a plurality of paper PDFs, each of which is an example of a document. That is, composite process extraction unit 111 extracts a composite process from the paper PDF. Then, composite process extraction unit 111 stores the extracted composite process in customer database 145 included in database group 140.

[0321] Note that combined process extraction unit 111 may be formed from any of the components included in combined process generation device 100 in the first to third embodiments above, other than document storage unit 102.

[0322] The extraction unit 112 downloads multiple pieces of paper data from the paper database 130. This causes the paper PDF and paper bib included in the paper data to be downloaded at the same time. Then, for each of the downloaded multiple pieces of paper data, the extraction unit 112 extracts each piece of information from the paper PDF and paper bib included in the paper data as customer information. Then, the extraction unit 112 stores a data set including the multiple pieces of customer information extracted from the paper data in the above-mentioned customer database 145.

[0323] Furthermore, when extracting the customer information from the paper data, the extraction unit 112 uses an equipment database 141, a material database 142, a project database 143, and a unit dictionary database 144 included in the database group 140.

[0324] Second analysis device 120 includes input unit 121 and estimation unit 122 .

[0325] Input unit 121 accepts various pieces of information in response to an input operation by a user of second analysis device 120. Such input unit 121 may be configured as, for example, a keyboard, a touch sensor, a touch pad, or a mouse.

[0326] The estimation unit 122 estimates various pieces of information relating to product purchases by customers, using a customer database 145, a device maintenance database 146, and a sales database 147, which will be described later.

[0327] The database group 140 includes an equipment database 141 , a material database 142 , a project database 143 , a unit dictionary database 144 , a customer database 145 , an equipment maintenance database 146 , a sales database 147 , and an estimation result database 148 .

[0328] The device database 141, also called a device DB, holds a plurality of pieces of information about each of a plurality of devices, such as the name, price, and specifications of the device.

[0329] The material database 142 is also called a material DB, and holds, for each of a plurality of materials, a plurality of pieces of information related to the material, such as the name and price of the material.

[0330] The project database 143, also called a project DB, holds a plurality of pieces of information about each of a plurality of projects, such as the name and budget of the project.

[0331] The unit dictionary database 144, also called a unit dictionary DB, holds information relating to units of weight.

[0332] The customer database 145 is also called a customer DB. In the customer database 145, as described above, a plurality of synthesis processes extracted by the synthesis process extraction unit 111 and a data set including a plurality of pieces of customer information extracted by the extraction unit 112 are stored.

[0333] The equipment maintenance database 146 is also called an equipment maintenance DB, and holds information on failures of equipment sold by sales companies. The information on equipment failures is, for example, a failure rate curve of the equipment. This failure rate curve shows an early failure period, an accidental failure period, and a wear-out failure period, and shows the time change of the failure rate in each period. The early failure period is a period in which defects in the design or manufacture of the equipment are likely to cause failures. The accidental failure period is a period in which the cause of failure is not related to the passage of time. The wear-out failure period is a period in which wear or wear over time is likely to cause failures. The estimation unit 122 uses such a failure rate curve to estimate the time-series change of the failure rate of the equipment. The above-mentioned failure rate curve held in the equipment maintenance database 146 is automatically or manually generated or updated, and is, for example, always kept in the latest state.

[0334] The sale database 147 is also called a sale DB, and holds sale conditions. The sale conditions are conditions relating to discounts on materials or equipment, for example, a condition that if Xg or more of material A is purchased at once, the price of the material A will be discounted. The estimation unit 122 may refer to such sale conditions to estimate various information relating to the purchase of products by customers. The sale conditions held in the sale database 147 are automatically or manually generated or updated, and are, for example, always kept up to date.

[0335] The estimation result database 148 is also called an estimation result DB. The estimation result database 148 stores information estimated by the estimation unit 122, such as material estimation information and device estimation information, which will be described later.

[0336] [Database and Extraction Unit 112 Processing] FIG. 38 is a diagram showing an example of a data table held in the device database 141. As shown in FIG.

[0337] The device database 141 holds, for example, a data table 141a shown in FIG. 38. This data table 141a shows, for each of a plurality of devices, a plurality of pieces of information related to the device. For example, the data table 141a shows, for each of a plurality of devices, the device ID, device name, device alias, device model number, sales company name, related device ID, price, start of use month, and specifications of the device in association with each other. The device ID is identification information for identifying the device, the device name is the name of the device, and the device alias is another name or abbreviation different from the name of the device. The device model number is the model number or model of the device determined by the sales company of the device. The sales company name is the name of the sales company of the device. The related device ID is identification information of a device related to a device having a device ID associated with the related device ID, for example, identification information of a device having the same use and function as the device having the device ID. The price is the price of the device determined by the sales company of the device having the device ID, and the start of use month is the month when the use of the device began. The month of start of use may be the month when the device was first sold. Specifications are specifications according to the use of the device. For example, the specifications of a device for ball milling include the ball diameter, the material of the device, or the disk rotation speed. On the other hand, the specifications of a device for annealing include the annealing method (e.g., far-infrared irradiation or hot air combined use), the maximum firing temperature, or the atmospheric conditions (nitrogen, argon, oxygen, etc.).

[0338] Specifically, data table 141a indicates, for a machine with machine ID "1", its machine ID "1", its machine name "Ball Milling Machine", its machine alias "BM", its machine model number "BZ204B", its distributor's name "Osaka Industries", its related machine ID "[3,4]", its machine price "80,000 yen", its machine start month of use "July 2011", and its machine specifications. Note that the related machine ID "[3,4]" indicates that the machines with machine IDs "3" and "4" are related to the machine with machine ID "1" as related machines.

[0339] In addition, even if multiple devices have the same device name and the same device model number, if the sales companies of the multiple devices are different, the prices of the multiple devices may also be different. Therefore, in data table 141a, even if multiple devices have the same device name and the same device model number, if the sales companies are different, each of the multiple devices is associated with a different device ID.

[0340] The data table 141a of such an equipment database 141 is a data table that can be created based on a catalog that each equipment sales company publishes on the company's web page or the like. This data table 141a may be created automatically or manually. In addition, since the lineup of equipment sold by equipment sales companies is continually updated, the data table 141a held in the equipment database 141 may always be updated to the latest state.

[0341] The extraction unit 112 in this embodiment extracts device information from paper data, for example, the paper PDF shown in (a) of Fig. 35, using BERT (Bidirectional Encoder Representations from Transformers). Note that BERT is a model based on deep learning that has been reported to have high performance in named entity extraction.

[0342] Then, the extraction unit 112 searches the data table 141a for a device name that matches the extracted device information. If the device name is not found, the extraction unit 112 searches the data table 141a for a device alias that matches the extracted device information. If the device alias is not found, the extraction unit 112 searches the data table 141a for a device model number that matches the extracted device information. When the extraction unit 112 finds a device name or the like by the search, the extraction unit 112 specifies the device ID, sales company name, related device ID, price, etc. associated with the device name or the like in the data table 141a as customer information related to the device. This allows the customer information related to the device to be extracted from the paper data. Then, the extraction unit 112 associates the paper ID of the paper data with the extracted customer information related to the device. The paper ID is identification information for identifying the paper data.

[0343] Here, there are variations in the notation of the device information described in each paper data. That is, the device information may be written as a formal name, an alias, an abbreviation, or a model number. In such a case, a problem may occur in which customer information cannot be appropriately extracted. Therefore, in this embodiment, as described above, by using the device alias and device model number in the data table 141a, the occurrence of problems due to the notation variations can be suppressed. For example, when the device information "BM II" is extracted from the paper data, the extraction unit 112 refers to each device alias shown in the data table 141a. Then, if the device alias matching the extracted device information is shown in the data table 141a, the extraction unit 112 specifies the device ID, etc. associated with the device alias in the data table 141a as customer information related to the device. As a result, even if the formal name of the device is, for example, "Ball Miling II", the occurrence of problems due to the notation variations of the names "Ball Miling II" and "BM II" can be suppressed.

[0344] If the data table 141a does not show a device name, device alias, or device model number that matches the extracted device information, the first analysis device 110 may update the data table 141a. That is, the first analysis device 110 updates the data table 141a so that the extracted device information is included in the data table 141a as a new device name. Specifically, the first analysis device 110 notifies the administrator of the device database 141 of the new device name by email or the like. This prompts the administrator to confirm the new device name, and further prompts the administrator to add the device model number, sales company name, price, and the like that correspond to the new device name. As a result, the data table 141a is updated, and the occurrence of problems due to variations in the notation of device information can be further suppressed.

[0345] In addition, the extraction unit 112 in this embodiment extracts device information from the paper data using BERT as described above, but may search for device names, device aliases, device model numbers, etc. shown in the data table 141a. Specifically, the extraction unit 112 searches the paper PDF for each device name shown in the data table 141a, and if no device name is found, searches the paper PDF for each device alias shown in the data table 141a. Then, if no device alias is found, the extraction unit 112 searches the paper PDF for each device model number shown in the data table 141a. When the extraction unit 112 finds the device name, device alias, or device model number, it specifies the device ID, sales company name, related device ID, price, etc. associated with the device name, etc. in the data table 141a as customer information related to the device. As a result, customer information related to the device is extracted from the paper data.

[0346] FIG. 39 is a diagram showing an example of a data table held in the material database 142. As shown in FIG.

[0347] The material database 142 holds, for example, a data table 142a shown in FIG. 39. This data table 142a shows, for each of a plurality of materials, a plurality of pieces of information related to the material. For example, the data table 142a shows, for each of a plurality of materials, the material ID, material name, material alias, sales company name, and price of the material in association with each other. The material ID is identification information for identifying the material, the material name is the composition formula of the material, and the material alias is a different name or abbreviation from the composition formula of the material. The sales company name is the name of the sales company of the material. The price is the price per gram of the material determined by the sales company of the material.

[0348] Specifically, for a material with material ID "3," data table 142a indicates the material ID "3," the material name of the material "LiPON," an alternative name for the material "Lithium phosphorus oxynitride," the name of the sales company for the material "Nakamura Material," and the price per gram of the material "1,500 yen."

[0349] In addition, even if multiple materials have the same material name, if the sales companies of the multiple materials are different, the prices of the multiple materials may also be different. Therefore, in the data table 142a, even if multiple materials have the same material name, if the sales companies are different, each of the multiple materials is associated with a different material ID.

[0350] The data table 142a of the material database 142 can be created based on a catalog published by each material sales company on the company's web page or the like. This data table 142a may be created automatically or manually. In addition, since the lineup of materials sold by material sales companies is continuously updated, the data table 142a held in the material database 142 may be constantly updated to the latest state.

[0351] The extraction unit 112 in this embodiment extracts material information from paper data, for example, the paper PDF shown in FIG. 35(b), using BERT in the same manner as described above. Then, the extraction unit 112 searches the data table 142a for a material name that matches the extracted material information. If the material name is not found, the extraction unit 112 searches the data table 142a for an alternative material name that matches the extracted material information. When the material name or the like is found by the search, the extraction unit 112 specifies the material ID, sales company name, price, and the like associated with the material name or the like in the data table 142a as customer information related to the material. This allows the customer information related to the material to be extracted from the paper data. Then, the extraction unit 112 associates the paper ID of the paper data with the customer information related to the extracted material.

[0352] Here, there are spelling variations in the material information described in each paper data. That is, the material information may include the official name, the alternative name, or the abbreviation. In such a case, there is a possibility that a problem of not being able to extract customer information appropriately may occur. Therefore, in this embodiment, as described above, by using the material alternative name in the data table 142a, it is possible to suppress the occurrence of problems due to the spelling variations. For example, when material information "Lithium phosphorus oxynitride" is extracted from the paper data, the extraction unit 112 refers to each material alternative name shown in the data table 142a. Then, if the material alternative name matching the extracted material information is shown in the data table 142a, the extraction unit 112 specifies the material ID, etc. associated with the material alternative name in the data table 142a as customer information. As a result, even if the official name of the material is, for example, "LiPON", it is possible to suppress the occurrence of problems due to spelling variations between the names "LiPON" and "Lithium phosphorus oxynitride".

[0353] If the data table 142a does not show a material name or a material alias that matches the extracted material information, the first analysis device 110 may update the data table 142a. That is, the first analysis device 110 updates the data table 142a so that the extracted material information is included in the data table 142a as a new material name. Specifically, the first analysis device 110 notifies the administrator of the material database 142 of the new material name by e-mail or the like. This prompts the administrator to confirm the new material name, and further prompts the administrator to add an equipment alias, a sales company name, a price, and the like that correspond to the new material name. As a result, the data table 142a is updated, and the occurrence of problems due to variations in the notation of material information can be further suppressed.

[0354] Furthermore, the extraction unit 112 in this embodiment extracts material information from the paper data using BERT as described above, but may also search for material names or material aliases shown in the data table 142a. Specifically, the extraction unit 112 searches the paper PDF for each material name shown in the data table 142a, and if no material name is found, searches the paper PDF for each material alias shown in the data table 142a. When the extraction unit 112 finds a material name or material alias, it specifies the material ID, sales company name, price, etc. associated with the material name, etc. in the data table 142a as customer information related to the material. As a result, customer information related to the material is extracted from the paper data.

[0355] FIG. 40 is a diagram showing an example of a data table held in the project database 143. As shown in FIG.

[0356] The project database 143 holds, for example, a data table 143a shown in FIG. 40. This data table 143a shows, for each of a plurality of projects, a plurality of pieces of information related to the project. For example, the data table 142a shows, for each of a plurality of projects, the project ID, project name, project alias, project identification number, start year, end year, and budget size of the project in association with each other. The project ID is identification information for identifying the project, the project name is the name of the project, and the project alias is a different name or abbreviation different from the name of the project. The project identification number is an identification number of the project determined by the issuer of the budget of the project. The issuer of the budget may be an organization that pays the budget or an organization that decides the budget. The start year is the year in which the project starts, determined by the issuer, and the end year is the year in which the project ends, determined by the issuer. The budget size is the budget of the project, determined by the issuer.

[0357] Specifically, for a project with project ID "1," data table 143a indicates the project ID "1," the project name of the project "Fund of material research," the alternative name of the project "MAT," the project identification number of the project "234-3344," the start year of the project "FY2018," the end year of the project "FY2020," and the budget size of the project "20 million yen."

[0358] The data table 143a of the project database 143 can be created based on information about each project, such as a national project, that is published on a web page or the like by a support organization that supports a research organization. The data table 143a may be created automatically or manually. In addition, since project information is updated continuously, the data table 143a stored in the project database 143 may always be updated to the latest state.

[0359] The extraction unit 112 in this embodiment extracts project information from paper data, for example, the paper PDF shown in FIG. 36, using BERT in the same manner as described above.

[0360] Here, the formats or patterns of the project identification numbers may differ among the multiple projects. Therefore, when the extraction unit 112 in this embodiment extracts project information, it determines whether the project information matches each pattern previously determined for the project identification number. For example, the project identification number patterns include a pattern in which "No." is placed immediately before a number, such as "No. 1234," a pattern in which numbers are separated by hyphens, such as "1233-3333," and a pattern in which alphabets and numbers are mixed, such as "AB1234." If the extraction unit 112 determines that the project information extracted from the paper matches any of the above patterns, it classifies the project information into a project identification number. On the other hand, if the extraction unit 112 determines that the project information extracted from the paper does not match any of the above patterns, it classifies the project information into a project name or an alternative project name.

[0361] Then, when the extracted project information is classified as a project name or a project alias, the extraction unit 112 searches the data table 143a for a project name that matches the project information. If the project name is not found, the extraction unit 112 searches the data table 143a for a project alias that matches the extracted project information. When the extraction unit 112 finds a project name or the like by the search, the extraction unit 112 specifies the project ID, start year, end year, budget size, and the like associated with the project name or the like in the data table 143a as customer information related to the project. Similarly, when the extracted project information is classified as a project identification number, the extraction unit 112 searches the data table 143a for a project identification number that matches the project information. When the extraction unit 112 finds a project identification number by the search, the extraction unit 112 specifies the project ID, start year, end year, budget size, and the like associated with the project identification number in the data table 143a as customer information related to the project. In this way, customer information related to the project is extracted from the paper data.

[0362] In addition, the extraction unit 112 in this embodiment extracts project information from the paper data using BERT as described above, but may search for the project name, project alias, project identification number, etc. shown in the data table 143a. Specifically, the extraction unit 112 searches the paper PDF for each project name shown in the data table 143a, and if no project name is found, searches the paper PDF for each project alias shown in the data table 143a. Then, if no project alias is found, the extraction unit 112 searches the paper PDF for each project identification number shown in the data table 143a. When the extraction unit 112 finds the project name, etc., it specifies the project ID, start year, end year, budget size, etc. associated with the project name, etc. in the data table 143a as customer information related to the project. As a result, customer information related to the project is extracted from the paper data.

[0363] FIG. 41 is a diagram showing an example of a data table held in the unit dictionary database 144. As shown in FIG.

[0364] The unit dictionary database 144 holds, for example, a data table 144a shown in FIG. 41. This data table 144a shows a plurality of unit symbols used for units of weight in association with the numerical notations corresponding to these unit symbols. The plurality of unit symbols are "g", "mg", and "μg". The numerical notation corresponding to the unit symbol "g" is "1", the numerical notation corresponding to the unit symbol "mg" is "0.001", and the numerical notation corresponding to the unit symbol "μg" is "0.000001". These numerical notations are used to convert weights expressed with the unit symbols "mg" and "μg" into weights expressed with the unit symbol "g", that is, grams.

[0365] The extraction unit 112 in this embodiment uses such data table 144a to extract the weight of the material expressed in the unit symbol "g" from the paper data, specifically the paper PDF. That is, the extraction unit 112 extracts the weight of the material as a gram amount from the paper data. Specifically, the extraction unit 112 searches the paper data for each of the multiple unit symbols shown in the data table 144a, and if any unit symbol is found, extracts the unit symbol and the numerical value written immediately before the unit symbol from the paper data. Then, the extraction unit 112 refers to the numerical notation associated with the extracted unit symbol in the data table 144a, and multiplies the numerical value of the numerical notation by the numerical value extracted from the paper PDF. In this way, the extraction unit 112 calculates the weight of the material expressed in the unit symbol "g", that is, the gram amount. As a result, the gram amount of the material is extracted from the paper PDF. For example, if the extraction unit 112 finds the unit symbol "mg" in the paper data, it extracts the unit symbol "mg" and the numerical value "1" written immediately before the unit symbol "mg" from the paper data. That is, "1mg" is extracted. Then, the extraction unit 112 refers to the numerical value notation "0.001" associated with the extracted unit symbol "mg" in the data table 144a. Next, the extraction unit 112 multiplies the numerical value of the numerical value notation "0.001" by the numerical value "1" to calculate the gram weight "0.001", which is the weight of the material expressed in the unit symbol "g". As a result, the gram weight "0.001g" is extracted from the paper PDF.

[0366] Moreover, the extraction unit 112 extracts customer information such as material ID from the paper data using the material database 142 as described above. Furthermore, the extraction unit 112 extracts the gram weight of a material from the paper data using the unit dictionary database 144. Then, the extraction unit 112 associates the gram weight and material ID extracted from each of the same paper data. Thus, the paper ID, material ID, and gram weight of the paper data are associated with each other. Moreover, the gram weight associated with such a material ID is treated as customer information, like the material ID.

[0367] When multiple material IDs and multiple gram amounts are extracted from the same research paper data, the material IDs and gram amounts may be associated with each other, i.e., paired, based on the positions in the research paper data where the material information and unit symbols are described. For example, the extraction unit 112 identifies the unit symbol and numerical value that are closest to the position where the material information corresponding to the material ID is described, and associates the gram amount corresponding to the unit symbol and numerical value with the material ID.

[0368] FIG. 42 is a diagram showing an example of an organization identification table and an author identification table.

[0369] The extraction unit 112 generates, for example, an organization identification table T1 shown in FIG. 42(a) and an author identification table T2 shown in FIG. 42(b) as auxiliary data of the customer database 145, and stores them in the customer database 145.

[0370] Specifically, when the extraction unit 112 extracts an organization name from, for example, a paper bib, the extraction unit 112 determines an organization ID for the organization name, associates the organization ID with the organization name, and includes the organization ID and the organization name in the organization identification table T1. The organization ID associated with the organization name may be determined, for example, according to the order in which the organization names are extracted.

[0371] The organization identification table T1 shows, for each of a plurality of organizations, an organization ID, which is identification information of the organization, in association with an organization name, which is the name of the organization. For example, the organization identification table T1 shows an organization ID "1" in association with the organization name "A University" of the organization having the organization ID "1".

[0372] In addition, when the extraction unit 112 extracts an author name from, for example, a paper bib, it determines an author ID for the author name, associates the author ID with the author name, and includes the associated author ID in the author identification table T2. The author ID associated with the author name may be determined, for example, according to the order in which the author names are extracted.

[0373] The author identification table T2 shows, for each of a plurality of authors, an author ID, which is the identification information of the author, in association with an author name, which is the name of the author. For example, the author identification table T2 shows, for the author with the author ID "1", the author ID "1" and the author name "K.Sasaki" in association with each other.

[0374] Scientific and technical papers are often written by multiple authors. In addition, there are cases where the authors belong to multiple organizations, or where the authors each belong to different organizations, such as a university and a company. In such cases, it is possible to search for published papers by each author and use the text information of the papers found to identify the organization (university, company, etc.) to which the author belongs.

[0375] In addition, when a student graduates from a university and gets a job as a researcher in a company, that student or researcher may publish papers as an author in a different organization. When a person moves between organizations, even if the author's name is the same between multiple papers, the author may not be able to be identified because the authors belong to different organizations.

[0376] Therefore, the extraction unit 112 may include a code for uniquely identifying a person, such as an ORCID (Open Researcher and Contributor ID), in association with the author name in the author identification table T2. This can solve the problem of an author being unable to be identified when an author with the same name publishes papers in multiple organizations. In addition, when an ORCID is included in the paper data in this embodiment, the extraction unit 112 may extract the ORCID from the paper data.

[0377] In recent years, a researcher may publish a list of published papers or a researcher's biography on the homepage or SNS (social networking service) of the researcher or the laboratory to which the researcher belongs. Therefore, a web search can be performed for the paper title extracted from the paper to find the homepage of the researcher who wrote the paper, and the researcher can be identified from the published paper list or biography published on the homepage. The author of a paper can also be identified by searching the paper title on a paper search site such as "Google (registered trademark) Scholar."

[0378] FIG. 43A is a diagram showing an example of a data table stored in the customer database 145. As shown in FIG.

[0379] The customer database 145 stores, for example, a data table 145a shown in FIG. 43A. This data table 145a includes a data set related to each of a plurality of paper data. The data set of the paper data indicates the paper ID, the author list, the organization list, the device-related information, the project-related information, the material-related information, and the publication month of the paper data in association with each other. The paper ID is identification information for identifying the paper data. The author list indicates, in a list format, the author ID corresponding to each of at least one author name described in the paper data. The organization list indicates, in a list format, the organization ID corresponding to each of at least one organization name described in the paper data. The device-related information indicates, in a dictionary format, a combination of the device ID of the device identified by the device information and the price of the device for each of at least one piece of device information described in the paper data. The project-related information indicates, in a dictionary format, a combination of the project ID of the project identified by the project information and the budget size of the project for each of at least one piece of project information described in the paper data. The material-related information indicates, for each of at least one piece of material information described in the paper data, a combination of the material ID of the material specified by the material information and the gram amount of the material in dictionary format. The publication month is the month in which the paper data was published or issued, and is the month that includes the publication date or issue date described above.

[0380] In this embodiment, when the extraction unit 112 downloads paper data from the paper database 130, it assigns a paper ID to the paper data. Then, the extraction unit 112 extracts the publication date or issue date from the paper bib of the paper data, and identifies the month including the publication date or issue date as the publication month. Furthermore, the extraction unit 112 uses the equipment database 141, the material database 142, the project database 143, and the unit dictionary database 144 to extract each customer information from the paper data as described above. Furthermore, when the extraction unit 112 extracts an organization name from the paper bib of the paper data, it identifies an organization ID associated with the organization name in the organization identification table T1. Similarly, when the extraction unit 112 extracts an author name from the paper bib of the paper data, it identifies an author ID associated with the author name in the author identification table T2.

[0381] The extraction unit 112 generates a dataset of the above-mentioned paper data using the paper ID assigned to the paper data, the organization ID and author ID identified for the paper data, each customer information extracted from the paper data, and the publication month identified from the paper data.

[0382] Specifically, the extraction unit 112 includes, for example, the paper ID "1" assigned to the paper data and the publication month "June 2019" identified from the paper data in the data set of the paper data. Furthermore, the extraction unit 112 extracts at least one author name from the paper bib of the paper data of the paper ID "1". For each of the at least one extracted author name, the extraction unit 112 selects an author ID associated with the author name in the author identification table T2 of (b) of FIG. 42. Then, the extraction unit 112 generates an author list, for example, [1,2], including the selected at least one author ID, and includes the author list [1,2] in the data set of the paper data. Note that this author list [1,2] includes the author ID "1" and the author ID "2".

[0383] Similarly, the extraction unit 112 extracts at least one organization name from the paper bib of the paper data with paper ID "1". For each of the at least one extracted organization name, the extraction unit 112 selects an organization ID associated with the organization name in the organization identification table T1 in FIG. 42(a). Then, the extraction unit 112 generates an organization list, for example, [1], including the at least one selected organization ID, and includes the organization list [1] in the dataset of the paper data. Note that this organization list [1] includes the organization ID "1".

[0384] Furthermore, the extraction unit 112 generates equipment-related information for the paper data using the customer information on the equipment extracted from the paper data as described above using the equipment database 141. For example, the extraction unit 112 generates equipment-related information {(1:2,000,000), (2:90,000)} by combining the equipment ID and price extracted as customer information on the equipment from the paper data of the paper ID "1". Then, the extraction unit 112 includes the generated equipment-related information {(1:2,000,000), (2:90,000)} in the data set of the paper data. Note that the equipment-related information {(1:2,000,000), (2:90,000)} includes a combination of the equipment ID "1" and the price "2 million yen" and a combination of the equipment ID "2" and the price "90,000 yen".

[0385] Similarly, the extraction unit 112 generates project-related information for the paper data using the customer information on the project extracted from the paper data as described above using the project database 143. For example, the extraction unit 112 generates project-related information {(2:20,000,000), (3:90,000,000)} by combining the project ID and the budget size extracted as customer information on the project from the paper data of the paper ID "1". Then, the extraction unit 112 includes the generated project-related information {(2:20,000,000), (3:90,000,000)} in the data set of the paper data. Note that the project-related information {(2:20,000,000), (3:90,000,000)} includes a combination of the project ID "2" and the budget size "20 million yen" and a combination of the project ID "2" and the price "90 million yen".

[0386] Similarly, the extraction unit 112 generates material-related information for the paper data using customer information on materials extracted from the paper data as described above using the material database 142. For example, the extraction unit 112 generates material-related information {2:0.1} by combining the material ID and gram amount extracted as customer information on materials from the paper data of the paper ID "1". Then, the extraction unit 112 includes the generated material-related information {2:0.1} in the data set of the paper data. Note that the material-related information {2:0.1} includes the combination of the material ID "2" and the gram amount "0.1g".

[0387] In this manner, the extraction unit 112 in this embodiment generates a data set of the paper data for each of the multiple downloaded paper data. Then, the extraction unit 112 stores the data sets in the customer database 145. As a result, a data table 145a consisting of the data sets is stored in the customer database 145. Note that the paper data is continuously published or issued. Therefore, the extraction unit 112 generates the data sets, for example, periodically, thereby keeping the data table 145a of the customer database 145 up to date.

[0388] FIG. 43B is a diagram showing another example of a data table stored in the customer database 145. As shown in FIG.

[0389] The customer database 145 may store a data table 145b shown in Fig. 43B. This data table 145b does not include the material-related information included in the data table 145a shown in Fig. 43A. In this respect, the data table 145b differs from the data table 145a, but otherwise the data table 145b shows the same information as the data table 145a.

[0390] For example, when the extraction unit 112 is unable to extract customer information related to materials from the paper data, the extraction unit 112 may generate a data set that does not include material-related information for the paper data and store the data set in the customer database 145. In this way, when customer information related to materials is not extracted from each of the multiple paper data, the data table 145b shown in Fig. 43B is stored in the customer database 145. Alternatively, the extraction unit 112 may store in the customer database 145 the data table 145b from which the material-related information is omitted for the purpose of the sales company of the device.

[0391] FIG. 43C is a diagram showing another example of a data table stored in the customer database 145. As shown in FIG.

[0392] The customer database 145 may store a data table 145c shown in Fig. 43C. This data table 145c does not include the device-related information included in the data table 145a shown in Fig. 43A. In this respect, the data table 145c differs from the data table 145a, but otherwise the data table 145c shows the same information as the data table 145a.

[0393] For example, when the extraction unit 112 is unable to extract customer information related to the equipment from the paper data, the extraction unit 112 may generate a data set that does not include equipment-related information for the paper data and store the data set in the customer database 145. As a result, when customer information related to the equipment is not extracted from each of the multiple paper data, the data table 145c shown in Fig. 43C is stored in the customer database 145. Alternatively, the extraction unit 112 may store the data table 145c, from which the equipment-related information is omitted, in the customer database 145 for the purpose of a material sales company.

[0394] Thus, in this embodiment, the support device 13 includes the first analysis device 110. The first analysis device 110 executes the following (b-1) to (b-3) for i=1 to n to generate a plurality of data sets, that is, data set 1 to data set n. The data set 1 to data set n are, for example, n data sets included in the data table 145a of the customer database 145 shown in FIG. 43A. Each of these data sets includes a paper ID and the like. In (b-1), the first analysis device 110 extracts information on at least one of the author of the document i and the organization to which the author belongs as the author-related information i from the document i. For example, the author-related information i corresponds to the author list and organization list in FIG. 43A. In (b-2), the first analysis device 110 extracts information on the device from the document i as the device-related information i. For example, the device-related information i corresponds to the device-related information in FIG. 43A. In addition, in (b-3), the first analysis device 110 generates a data set i showing the author-related information i and the device-related information i in association with each other. That is, a data set including the above-mentioned author list, organization list, and device-related information is generated.

[0395] As a result, data sets 1 to n are generated from documents 1 to n. Data set i indicates the author-related information i and device-related information i of document i in association with each other, so by looking at data sets 1 to n, it is easy to understand which author or organization is using which device. This can improve the efficiency of sales activities by device sales companies.

[0396] In this embodiment, in generating the multiple data sets, the first analysis device 110 further executes the following (b-4) with i=1 to n. That is, in (b-4), the first analysis device 110 extracts information on the type of material from the document i as material information i. Then, in (b-3), the first analysis device 110 generates a data set i that indicates the material information i in further association with the author-related information i. For example, the material information i may be a material ID included in the material-related information in FIG. 43A.

[0397] As a result, the data set i shows the author-related information i and material information i of the document i in association with each other, so by looking at the data sets 1 to n, it is easy to understand which author or organization is using what material. This makes it possible to improve the efficiency of sales activities by material sales companies.

[0398] In this embodiment, the first analysis device 110 further executes the following (b-5) for i=1 to n in generating the multiple data sets. That is, in (b-5), the first analysis device 110 extracts information on the amount of material from the document i as amount information i. Then, in (b-3), the first analysis device 110 generates a data set i that indicates the amount information i in further association with the author-related information i. For example, the amount information i may be the amount in grams included in the material-related information in FIG. 43A.

[0399] As a result, the data set i shows the author-related information i, material information i, and quantity information i of the document i in association with each other. Therefore, by looking at the data sets 1 to n, it is easy to understand which author or organization is using what material and how much of it. This can further improve the efficiency of sales activities by material sales companies.

[0400] In this embodiment, in generating the multiple data sets, the first analysis device 110 further executes the following (b-6) with i=1 to n. That is, in (b-6), the first analysis device 110 extracts information on a project supporting an author or organization from the document i as project-related information i. Then, in (b-3), the first analysis device 110 generates a data set i that indicates the project-related information i in further association with the author-related information i. For example, the project-related information i may be the project-related information in FIG. 43A.

[0401] As a result, the data set i shows the author-related information i and project-related information of the document i in association with each other. Therefore, by looking at the data sets 1 to n, it is easy to understand which author or organization is receiving support from which project. Therefore, it is possible to estimate the repayment capacity of the author or organization, and to further improve the efficiency of sales activities to the author or organization.

[0402] [Processing of the second analyzer] The estimation unit 122 of the second analysis device 120 acquires the data table 145a from the customer database 145, and converts the data table 145a into an organization data table.

[0403] FIG. 44 is a diagram illustrating an example of an organization data table.

[0404] The estimation unit 122 converts the data table 145a of the customer database 145 into, for example, an organization data table 145x shown in FIG. 44 by sorting each data set included in the data table 145a by organization ID.

[0405] As shown in FIG. 44, the organization data table 145x indicates, for each of a plurality of organizations, the organization ID, device-related information, project-related information, material-related information, and paper author information of the organization in association with each other.

[0406] When converting the data table 145a of the customer database 145 into the organization data table 145x, the estimation unit 122 refers to a plurality of organization lists included in the data table 145a. Then, the estimation unit 122 selects a plurality of organization IDs from the organization lists, for example, in ascending order. The estimation unit 122 generates one new device-related information in a dictionary format by combining all device-related information associated with the selected organization ID in the data table 145a. For example, the estimation unit 122 generates one new device-related information {(2:400,000), (3:4,000,000), (4:8,000,000), (5:12,000,000), (6:400,000)} by combining all device-related information associated with the organization ID "2" in the data table 145a.

[0407] Also, there are cases where the same device ID associated with the same organization ID is included in each of the multiple data sets in the data table 145a. In other words, there are cases where the same device and its price are described in each of the multiple paper data published by the same organization. In such a case, the estimation unit 122 calculates the sum of the prices corresponding to the same device ID included in those data sets, that is, the sum of the prices of the same device described in each of the multiple papers. For example, assume that in the data table 145a, a data set having an organization ID "2" and a paper ID "1" and a data set having the same organization ID "2" and another paper ID "2" each contain {1:1,000,000} as device-related information. That is, the paper data of the paper ID "1" and the paper ID "2" published by the organization ID "2" each contain a device with device ID "1" and a price of the device "1 million yen". In such a case, the estimation unit 122 calculates the total price "1 million yen + 1 million yen = 2 million yen" of the price "1 million yen" corresponding to the device ID "1" included in those data sets. As a result, the estimation unit 122 generates one new piece of apparatus-related information {1:2,000,000}.

[0408] In this way, the total price of the equipment used in the experiments in a paper is calculated for each organization ID, making it possible to understand what type and how much of each equipment is purchased by each organization. For example, if the total price of equipment A in organization X exceeds the total price of equipment B, it can be interpreted that the demand for equipment A in organization X is higher than that of equipment B.

[0409] Next, the estimation unit 122 generates one new project-related information in a dictionary format by combining all project-related information associated with the selected organization ID in the data table 145a. For example, the estimation unit 122 generates one new project-related information {(2:400,000), (3:4,000,000), (4:8,000,000), (5:12,000,000), (6:400,000)} by combining all project-related information associated with the organization ID "2" in the data table 145a.

[0410] Next, the estimation unit 122 generates one new material-related information in a dictionary format by combining all the material-related information associated with the selected organization ID in the data table 145a. For example, the estimation unit 122 generates one new material-related information {1:0.1, 2:0.4, 3:0.7} by combining all the material-related information associated with the organization ID "2" in the data table 145a.

[0411] Also, each of the multiple data sets in the data table 145a may contain the same material ID associated with the same organization ID. In other words, the same material and its weight may be described in each of the multiple paper data published by the same organization. In such a case, the estimation unit 122 calculates the total gram amount corresponding to the same material ID included in those data sets, that is, the total weight of the same material described in each of the multiple papers. For example, assume that in the data table 145a, a data set having an organization ID "2" and a paper ID "1" and a data set having the same organization ID "2" and another paper ID "2" each contain {1:0.1} and {1:0.2} as material related information. That is, the paper data of the paper ID "1" published by the organization with the organization ID "2" describes a material with the material ID "1" (e.g., LiPON) and the weight of the material "0.1g". Furthermore, another paper data with paper ID "2" issued by the organization with the same organization ID "2" describes a material with material ID "1" (e.g. LiPON) and its weight "0.2g". In this case, the estimation unit 122 calculates the total gram weight "0.1g + 0.2g = 0.3g" of the gram weights "0.1g" and "0.2g" corresponding to the material ID "1" included in those data sets. As a result, the estimation unit 122 generates one new piece of material-related information {1:0.3}.

[0412] In addition, in the experimental data in scientific and technical papers, the reproducibility of the experiment is emphasized. In the experiment of the synthesis process, the results of multiple experiments are often published in the paper data, not the results of one experiment. Therefore, in the support device 13, for the amount of material used in the experiment of the synthesis process in grams, the amount of material in grams and the number of experiments may be extracted from the paper data, and the product of these may be calculated as the amount of material in grams used in the experiment of the synthesis process. In addition, if the number of experiments is not stated in the paper data, the number of experiments may be estimated using the number of points plotted in a graph or table, etc.

[0413] In this way, the total amount in grams of materials used in experiments in a paper is calculated for each organization ID, making it possible to grasp what type and how much of each material is purchased by each organization. For example, if the total amount in grams of material A in organization X exceeds the total amount in grams of material B, it can be interpreted that the demand for material A is higher than that of material B in organization X.

[0414] Next, the estimation unit 122 generates paper author information by combining all pairs of paper IDs and author lists associated with the selected organization ID in the data table 145a. For example, the estimation unit 122 generates paper author information {2:[2,3],3:[3]} by combining all pairs of paper IDs and author lists associated with the organization ID "2" in the data table 145a. Note that the paper author information includes a pair of paper ID "2" and author list [2,3], and a pair of paper ID "3" and author list [3].

[0415] Then, the estimation unit 122 associates the device-related information, project-related information, material-related information, and paper author information generated for the selected organization ID as described above with the organization ID. By performing such association for each of the multiple organization IDs, an organization data table 145x shown in FIG. 44 is generated.

[0416] [Estimation by the estimation part] The estimation unit 122 estimates, as material estimated information, information indicating the material purchase prospect, the repayment capacity, and the profit margin of each of the multiple organizations, for example, using the organization data table 145x shown in Fig. 44. Similarly, the estimation unit 122 estimates, as equipment estimated information, information indicating the equipment purchase prospect and the repayment capacity of each of the multiple organizations, for example, using the organization data table 145x shown in Fig. 44.

[0417] The likelihood of purchasing a material is a value that indicates the strength of the organization's willingness to purchase a material as a customer. Similarly, the likelihood of purchasing a device is a value that indicates the strength of the organization's willingness to purchase a device as a customer.

[0418] The profit margin for a material is the ratio between the price per gram of material and the profit included in that price.

[0419] The ability to pay for materials and the ability to pay for equipment are indicators of the financial status of the organization, which is the customer. By estimating such ability to pay, the material or equipment sales company can understand the financial status of the organization, which is the customer. As a result, the sales company can narrow down the types of materials or equipment to be sold and thereby improve the efficiency of sales activities.

[0420] FIG. 45A is a diagram showing an example of a result of estimation by the estimation unit 122 regarding materials.

[0421] As described above, the estimation unit 122 estimates the material estimation information of each of the multiple tissues by using, for example, the tissue data table 145x shown in FIG.

[0422] For example, the estimation unit 122 extracts a plurality of pieces of extraction information shown in the organization data table 145x. Each of the plurality of pieces of extraction information is a combination of an organization ID and a material ID included in the material related information associated with the organization ID. Then, for each piece of extraction information, the estimation unit 122 estimates material estimation information for the material of the material ID included in the extraction information, in the organization of the organization ID included in the extraction information. For example, as shown in FIG. 45A, the estimation unit 122 estimates the purchase likelihood, repayment ability, and profit margin for the material of the material ID "3" in the organization of the organization ID "1". Note that the combination of the organization ID "1" and the material ID "3" is the extraction information, and the estimated purchase likelihood, repayment ability, and profit margin are included in the material estimation information.

[0423] In estimating the likely purchase of a material, the estimation unit 122 converts the total of the gram amounts associated with the extraction information of the estimation target in the texture data table 145x into a total price, and estimates the total price as the likely purchase.

[0424] In a specific example of a purchase likelihood estimation, the extracted information to be estimated is a combination of organization ID "1" and material ID "2". In this case, the estimation unit 122 derives the total gram amount of the material with material ID "2" used in organization ID "1" from the organization data table 145x. In the example of FIG. 44, the estimation unit 122 derives the total gram amount of "0.1 g". Then, the estimation unit 122 refers to the price per gram of the material "4000 yen" associated with the material ID "2" in the data table 142a of the material database 142. The estimation unit 122 multiplies the price per gram of the material "4000 yen" by the derived total gram amount of "0.1 g" to derive the total price of the material with material ID "2" of "400 yen". As a result, the above-mentioned total price of "400 yen" is estimated as the purchase likelihood of the material with material ID "2" in the organization with organization ID "1".

[0425] Here, the estimation unit 122 may convert the purchase likelihood expressed by the numerical value that is the total price into a purchase likelihood expressed by any one of three levels of "small, medium, high". For example, the estimation unit 122 acquires a first threshold value and a second threshold value of the purchase likelihood from the input unit 121 in response to an input operation to the input unit 121 by a user of the second analysis device 120. Then, if the total price is less than the first threshold value, the estimation unit 122 converts the purchase likelihood expressed by the total price into a purchase likelihood expressed by a level "small". Also, if the total price is equal to or greater than the first threshold value and less than the second threshold value, the estimation unit 122 converts the purchase likelihood expressed by the total price into a purchase likelihood expressed by a level "medium". Then, if the total price is equal to or greater than the second threshold value, the estimation unit 122 converts the purchase likelihood expressed by the total price into a purchase likelihood expressed by a level "large".

[0426] Next, in estimating the solvency of materials, the estimation unit 122 estimates the total amount of the budget scale of the project-related information associated with the extracted information of the estimation target in the organization data table 145x as the solvency. In other words, the estimation unit 122 estimates the total amount of the budget scale of the project-related information associated with the organization ID included in the extracted information of the estimation target as the solvency.

[0427] In a specific example of repayment ability estimation, the extracted information of the estimation target includes organization ID "1". In this case, the estimation unit 122 derives project-related information of organization ID "1" from the organization data table 145x. In the example of FIG. 44, the estimation unit 122 derives project-related information {(2:20,000,000), (3:90,000,000)}. Then, the estimation unit 122 estimates the total amount of the budget scales to be "110 million yen" as the repayment ability of the organization with organization ID "1".

[0428] In addition, as described above, the estimation unit 122 may convert the ability to pay expressed by a numerical value that is the total price into an ability to pay expressed at any one of three levels: "low, medium, high."

[0429] Next, in estimating the profit margin of the material, the estimation unit 122 estimates the ratio between the price per 1g of the material of the material ID included in the extraction information of the estimation target and the profit included in the price as the profit margin. The profit is an amount obtained by subtracting the cost per 1g of the material from the price per 1g of the material. In other words, the estimation unit 122 estimates the profit margin for the material of the material ID included in the extraction information in the organization of the organization ID included in the extraction information as the above-mentioned ratio. Specifically, the estimation unit 122 refers to the data table 142a of the material database 142, and derives the price per 1g of the material of the material ID included in the extraction information of the estimation target from the data table 142a. Furthermore, the estimation unit 122 acquires the cost per 1g of the material of the material ID from the input unit 121 in response to an input operation to the input unit 121 by the user of the second analysis device 120. Then, the estimation unit 122 calculates the profit margin by dividing the difference between the price per gram of the material and the cost price by the price per gram of the material.

[0430] In a specific example of profit margin estimation, the extraction information to be estimated includes a material ID "2". In this case, the estimation unit 122 derives the price per 1g of the material with the material ID "2" of "4000 yen" from the data table 142a of the material database 142. Furthermore, the estimation unit 122 acquires the cost per 1g of the material with the material ID of "2" from the input unit 121 in response to an input operation by a user of the second analysis device 120 to the input unit 121. The estimation unit 122 then calculates a profit margin of "0.5" by dividing the difference "2000 yen" between the price per 1g of the material "4000 yen" and the cost "2000 yen" by the price per 1g of the material "4000 yen". As a result, the profit margin of "0.5" for the material with the material ID "2" in the organization with the organization ID "1" is estimated to be "0.5".

[0431] Furthermore, the estimation unit 122 may convert the profit margin expressed by the ratio into a profit margin expressed at one of three levels: "small, medium, high," as described above.

[0432] Furthermore, when estimating the profit margin, the estimation unit 122 may refer to the sale conditions stored in the sale database 147 and estimate the profit margin based on the sale conditions. For example, if the total price indicated by the purchase likelihood satisfies the sale conditions, the estimation unit 122 discounts the price per gram of material. Specifically, the estimation unit 122 subtracts a predetermined discount rate (e.g., 10%) from the price. Then, the estimation unit 122 estimates the profit margin based on the discounted price.

[0433] The estimation unit 122 stores material estimation information including the estimated purchase prospect, payment ability, and profit margin of the material in the estimation result database 148 included in the database group 140. Furthermore, the estimation unit 122 may output the material estimation information to a presentation device such as a display, and present the material estimation information to the user of the second analysis device 120 in a table format as shown in the example of FIG. 45A. This allows the user to understand which organization is likely to purchase which material in what quantity by referring to the material estimation information. Furthermore, the user can observe from a bird's-eye view for each organization what profit margin can be expected if the material is sold.

[0434] Furthermore, the estimation unit 122 may acquire a material ID from the input unit 121 in response to an input operation to the input unit 121 by a user of the second analysis device 120. In this case, the estimation unit 122 may present to the user only the material estimation information associated with the material ID. This allows only the material estimation information related to the material sold by the material sales company to which the user belongs to be presented to the user. This makes it possible to suppress presentation of information unnecessary for the material sales company, and to make necessary information easy to view for the user.

[0435] FIG. 45B is a diagram showing another example of the estimation result related to the material by the estimation unit 122.

[0436] The estimation unit 122 may reflect the Web search result on the purchase likelihood of the above-mentioned material. In this case, the estimation unit 122 estimates material estimation information including the purchase likelihood, the Web search reflection likelihood, the payment ability, and the profit rate for each of the multiple extracted information, as shown in FIG. 45B.

[0437] For example, when the estimation unit 122 estimates the purchase likelihood of an organization having an organization ID included in the extracted information as described above, the estimation unit 122 identifies the name of a sales company of a material used by the organization by referring to the data table 142a of the material database 142. That is, the estimation unit 122 identifies the name of a sales company associated with the material ID included in the extracted information in the data table 142a. Then, the estimation unit 122 acquires the number of views of a homepage published by the company having that sales company name. The estimation unit 122 calculates a view ratio by dividing the acquired number of views by a predetermined reference number of views. Then, the estimation unit 122 calculates a web search reflection likelihood by multiplying the purchase likelihood expressed by the total price as described above by the view ratio. Note that, like the purchase likelihood described above, the estimation unit 122 may convert the web search reflection likelihood expressed by the numerical value into a web search reflection likelihood expressed by one of three levels of "low, medium, high".

[0438] Such a web search reflection probability reflects the number of times the homepage of the material sales company is viewed. The number of times the homepage is viewed by an organization such as a university or research institute to purchase the material is included in the number of times the homepage is viewed. Therefore, the web search reflection probability may be more accurate than a purchase probability that does not reflect the web search results. Therefore, a user of the second analysis device 120 can more appropriately grasp the purchase probability of each customer by referring to the web search reflection probability.

[0439] FIG. 46 is a diagram showing an example of displaying the result of estimation regarding materials by the estimation unit 122.

[0440] The estimation unit 122 may display the purchase likelihood and profit margin of each organization on a display as a graph, for example, as shown in Fig. 46. For example, the horizontal axis of the graph indicates the profit margin, and the vertical axis indicates the purchase likelihood. In such a graph, the purchase likelihood and profit margin of each organization, such as A research institute, P Corporation, T University, and W University, are expressed by the position of a point.

[0441] By observing such a graph, the user of the second analysis device 120 can observe the marketing of each organization in a multi-dimensional manner. For example, in the graph shown in FIG. 46, it is possible to intuitively grasp that T University has the highest purchase prospect. As a result, if the user of the second analysis device 120 belongs to a material sales company, the user can determine that T University has the highest sales efficiency among T University, P Corporation, W University, and A research institute.

[0442] In the example of Fig. 46, the estimation unit 122 displays a graph showing purchase likelihood and profit margin, but may generate and display a graph showing purchase likelihood and repayment ability, or a graph showing profit margin and repayment ability. The estimation unit 122 may generate and display a three-axis graph showing each of purchase likelihood, repayment ability, and profit margin, instead of a two-axis graph as in Fig. 46.

[0443] FIG. 47 is a diagram showing another example of the result of estimation by the estimation unit 122 regarding the material.

[0444] The estimation unit 122 may use the data table 145a of the customer database 145 to derive a time series change in the likelihood of purchasing the material to be processed in the organization to be processed, as shown in Fig. 47, and display it on a display. For example, the estimation unit 122 identifies at least one combination of gram amount and publication month associated with organization ID "1" and material ID "2" in the data table 145a. Note that organization ID "1" is an example of an organization ID for identifying the organization to be processed, and material ID "2" is an example of a material ID for identifying the material to be processed.

[0445] For example, in the data table 145a shown in FIG. 43A, the estimation unit 122 identifies a combination of the gram amount "0.1 g" and the publication month "June 2019". Furthermore, the estimation unit 122 identifies the price per gram of the material associated with the material ID "2" of "4000 yen" by referring to the data table 142a of the material database 142. Then, the estimation unit 122 multiplies the gram amount "0.1 g" identified as described above by the price "4000 yen" to derive the payment price "400 yen" as the purchase prospect of the material with the material ID "2" in the organization with the organization ID "1". As a result, "400 yen" is derived as the purchase prospect of the material with the material ID "2" in the organization with the organization ID "1" in the publication month "June 2019". The estimation unit 122 derives such a purchase likelihood for each of at least one combination identified as described above, thereby deriving a time series change in the purchase likelihood of the material with material ID "2" in the organization with organization ID "1".

[0446] In this manner, in this embodiment, the time series change in the annual purchase prospect of a certain material in a certain organization can be visualized. By reading the maximum points of the purchase prospect from this time series change, the period in which sales activities are carried out can be clarified. For example, in the example of FIG. 47, the maximum points of the purchase prospect are March, June, and November 2019. Therefore, it is found that it is good to carry out sales activities in March, June, and November every year, and sales efficiency can be improved.

[0447] FIG. 48 is a diagram showing an example of the result of estimation regarding the device by the estimation unit 122. As shown in FIG.

[0448] As described above, the estimation unit 122 estimates the device inferred information of each of the multiple organizations by using, for example, the organization data table 145x illustrated in FIG.

[0449] For example, the estimation unit 122 extracts a plurality of pieces of extracted information shown in the organization data table 145x. Each of the plurality of pieces of extracted information is a combination of an organization ID and a device ID included in the device related information associated with the organization ID. The estimation unit 122 may also include an associated device ID corresponding to the device ID in the extracted information. That is, the estimation unit 122 may refer to the data table 141a of the device database 141, and include an associated device ID associated with the above-mentioned device ID in the data table 141a in the extracted information.

[0450] Then, for each piece of extraction information, the estimation unit 122 estimates device estimated information for a device with a device ID included in the extraction information, in an organization with an organization ID included in the extraction information. For example, as shown in Fig. 48, the estimation unit 122 estimates purchase likelihood and repayment ability for a device with a device ID "1" in an organization with an organization ID "1". Note that the combination of organization ID "1" and device ID "1" is included in the extraction information, and the estimated purchase likelihood and repayment ability are included in the device estimated information.

[0451] In estimating the likelihood of purchasing a device, the estimation unit 122 estimates the total price of devices associated with the extraction information of the estimation target in the organization data table 145x as the likelihood of purchasing the device.

[0452] In a specific example of a purchase likelihood estimation, the extracted information of the estimation target includes a combination of organization ID "1" and device ID "1". In this case, the estimation unit 122 derives the price of the device with device ID "1" used in the organization with organization ID "1" from the organization data table 145x. In the example of FIG. 44, the estimation unit 122 derives the device price of "2 million yen". As a result, the above-mentioned price "2 million yen" is estimated as the purchase likelihood of the device with device ID "1" in the organization with organization ID "1".

[0453] In addition, as described above, the estimation unit 122 may convert the purchase likelihood expressed by a numerical value that is the price into a purchase likelihood expressed at any one of three levels: "low, medium, high."

[0454] In estimating the solvency of the equipment, the estimation unit 122 estimates the total amount of the budget scale of the project-related information associated with the extracted information of the estimation target in the organization data table 145x as the solvency. In other words, the estimation unit 122 estimates the total amount of the budget scale of the project-related information associated with the organization ID included in the extracted information of the estimation target as the solvency. The solvency of this equipment is estimated in the same way as the solvency of materials.

[0455] The estimation unit 122 stores the estimated device purchase prospect and the device estimation information including the repayment capacity of the device in the estimation result database 148 included in the database group 140. Furthermore, the estimation unit 122 may output the device estimation information to a presentation device such as a display, and present the device estimation information to the user of the second analysis device 120 in a table format as shown in the example of FIG. 48. This allows the user to understand how much of which device is likely to be purchased by which organization by referring to the device estimation information. In addition, the user can get an overview of the purchase prospect for each organization, and furthermore, by checking the repayment capacity, it becomes possible to estimate the success probability of sales. Furthermore, the device ID related to the device ID is also displayed at the same time. This makes it possible to flexibly guide the organization according to its financial situation, such as introducing a new model of the device currently used by the organization, or introducing a cheaper version of the device to an organization with low repayment capacity.

[0456] Furthermore, the estimation unit 122 may acquire a device ID from the input unit 121 in response to an input operation to the input unit 121 by a user of the second analysis device 120. In this case, the estimation unit 122 may present to the user only device estimation information associated with the device ID. This allows only device estimation information related to devices sold by the device sales company to which the user belongs to be presented to the user. This makes it possible to suppress presentation of information unnecessary for the device sales company, and to make necessary information easy to view for the user.

[0457] FIG. 49 is a diagram showing another example of the result of estimation regarding the device by the estimation unit 122. In FIG.

[0458] The estimation unit 122 may estimate a time series change in the failure rate of the device as shown in FIG. 49. For example, the estimation unit 122 refers to the data table 145a of the customer database 145, and identifies the organization ID and publication month associated with the device ID of the device to be estimated. For example, if the device ID of the device to be estimated is "1", the estimation unit 122 identifies the organization ID and publication month associated with the device-related information including the device ID "1". When the data table 145a shown in FIG. 43A is referred to, the organization ID "1" and the publication month "June 2019" are identified.

[0459] Next, the estimation unit 122 obtains a failure rate curve associated with the equipment ID of the equipment to be estimated from the equipment maintenance database 146. This failure rate curve indicates an early failure period, an accidental failure period, and a wear-out failure period, and indicates the change in failure rate over time in each of these periods. For example, the early failure period is three months from the start of use of the equipment, the accidental failure period is five years after the early failure period has elapsed, and the wear-out failure period is the period after the accidental failure period has elapsed.

[0460] The estimation unit 122 sets the start month of the change in the failure rate indicated by the acquired failure rate curve to the publication month identified as described above. For example, if the publication month is "June 2019", the start month of the change in the failure rate is set to "June 2019". As a result, the estimation unit 122 estimates the failure rate curve in which the failure rate changes from the start month "June 2019" as a time-series change in the failure rate in the device with device ID "1". Therefore, the estimation unit 122 determines that the period from June 2019 to September 2019 is the early failure period, the period from September 2019 to September 2024 is the random failure period, and the period after September 2024 is the wear-out failure period.

[0461] The estimation unit 122 stores failure rate information including information indicating the time-series change in the failure rate of the device estimated in this manner, the device ID of the device, and the organization ID specified as described above, in the estimation result database 148. Furthermore, the estimation unit 122 may output the failure rate information to a presentation device such as a display, and present the time-series change in the failure rate of the device to the user of the second analysis device 120 in a graph format as shown in the example of FIG.

[0462] This allows the user of second analysis device 120 to intuitively know when to notify the user of device maintenance. During the early failure period and wear-out failure period, when the device failure rate is high, the user may be notified of maintenance.

[0463] In addition, scientific and technical papers published in academic journals include the date the paper was submitted, the date the reviewer decided to accept it, and the date it was published in the journal. In particular, in prestigious journals, it can take more than a year from the time the author submits the paper until it is published. Also, the device described in the paper was purchased before the paper was submitted. Therefore, the date the paper was submitted can be used instead of the publication month mentioned above.

[0464] In recent years, experimental results and the like may be made publicly available as archives on the Internet. Furthermore, similar experimental results may be presented at international conferences, etc. Therefore, for papers presented at international conferences, etc., you may search the archives on the Internet, and if experimental results similar to those of the presented papers are published, you may use the date of the archive instead of the publication month mentioned above.

[0465] In addition, the name of the same device may appear in multiple scientific and technical papers written by the same author. However, since the devices used in experiments on material synthesis processes are expensive, it is rare for multiple units to be owned. Therefore, if the name of the same device is extracted from multiple paper data, the publication date of the oldest paper data may be used instead of the publication month mentioned above.

[0466] [Processing flow] FIG. 50A is a flowchart showing an example of processing by extraction section 112.

[0467] The extraction unit 112 selects and downloads one piece of paper data from the multiple pieces of paper data held in the paper database 130 (step S51). Next, the extraction unit 112 extracts the publication month of the paper data from the paper bib of the paper data (step S52). That is, the extraction unit 112 extracts the publication date or publication date of the paper data and identifies the publication month including the publication date or publication date. Furthermore, the extraction unit 112 extracts the author name as an author ID from the paper data (step S53) and the organization name as an organization ID (step S54). The author ID may be an ORCID. The extraction unit 112 may also check whether the extracted author name belongs to the organization of the extracted organization name. For example, the extraction unit 112 may search for past documents of the extracted author name and compare the extracted author name with the organization name by referring to the data of the documents.

[0468] Next, the extraction unit 112 extracts material-related information (step S55), project-related information (step S56), and equipment-related information (step S57) from the paper data. In extracting these pieces of information, data table 142a of material database 142, data table 143a of project database 143, and data table 141a of equipment database 141 are used. Even if there are variations in the notation of material information, project information, and equipment information in the paper data, the extraction unit 112 can appropriately extract material-related information, project-related information, and equipment-related information by referring to the above data tables.

[0469] Then, the extraction unit 112 stores a data set including each piece of information extracted in steps S52 to S57 in the customer database 145 in association with the paper ID of the paper data selected in step S51 (step S58).

[0470] After step S58, the extraction unit 112 determines whether or not all of the paper data held in the paper database 130 has been selected (step S59). If the extraction unit 112 determines that no selection has been made (No in step S59), it repeats the process from step S51. On the other hand, if the extraction unit 112 determines that selection has been made (Yes in step S59), it ends the process.

[0471] FIG. 50B is a flowchart showing an example of the process of the estimation unit 122.

[0472] The estimation unit 122 acquires a data table 145a from the customer database 145 (step S61). Next, the estimation unit 122 converts the data table 145a into an organization data table 145x by sorting each data set included in the data table 145a by organization ID (step S62).

[0473] Next, the estimation unit 122 estimates material estimation information (step S63) and further estimates equipment estimation information (step S64) using the texture data table 145x. Then, the estimation unit 122 estimates a time series change in the equipment failure rate (step S65).

[0474] The support device 13 in this embodiment can extract customer information efficiently by focusing on the information medium of paper data that is reliably and continuously published by material researchers, etc. Furthermore, by analyzing the extracted customer information, it is possible to provide information that is useful for the sales activities of material or device sales companies.

[0475] In this embodiment, paper data such as science and technology papers are used, but reports or final reports of national projects may also be used. In reports of national projects, information on the devices and materials purchased as the results is often described. In particular, unlike science and technology papers, the number of pages is often not limited in such reports, and many contents are described. On the other hand, in science and technology papers, the number of pages is often limited in international conference academic journals. Therefore, when the project name, etc. is described in paper data such as science and technology papers, it is possible to collect information on the devices and materials purchased for experiments, etc. of the paper data by searching text information such as the final report of the project. However, since national projects often span multiple years, the final report of a national project, etc. is often made public after the end year of the national project. On the other hand, science and technology papers published in international conferences or technical journals are often made public earlier than that.

[0476] Therefore, in this embodiment, the extraction unit 112 first constructs the device-related information in the customer database 145 from published paper data. Then, when the experimental results shown in the paper data are the research results of a project, the extraction unit 112 may update the device-related information using a report or the like that is published after the project is completed.

[0477] Further, in this embodiment, the paper data is used, but other documents may be used instead of the paper data.

[0478] Furthermore, first analysis device 110 in the present embodiment includes synthesis process extraction unit 111, but may not include synthesis process extraction unit 111. In this case, the paper data stored in paper database 130 may include content other than the synthesis process.

[0479] In this embodiment, the publication date or the publication date may be extracted instead of the publication month and shown in the data table 145a of the customer database 145.

[0480] Although the support device, generation device, and analysis device according to one or more aspects have been described based on the above-mentioned embodiments and modifications, the present disclosure is not limited to these embodiments and modifications. As long as it does not deviate from the gist of the present disclosure, various modifications conceived by a person skilled in the art to the above-mentioned embodiments and modifications, and forms constructed by combining components in different embodiments or different modifications may also be included within the scope of the present disclosure.

[0481] For example, in the above embodiments and modifications, the document is a paper, but it may be a textbook, a magazine, a patent document, or the like.

[0482] In the above first to third embodiments and the modified examples, the composite process generated by composite process generation device 100 is stored in composite process storage unit 201, but the document used in generating the composite process may also be stored in association with the composite process. In this case, the composite process or the composite composite process can be displayed on display unit 205, and the document that is the source of the composite process can also be displayed on display unit 205 in response to a user operation. Instead of storing the document itself in composite process storage unit 201, information indicating the location where the document is saved may be stored in composite process storage unit 201. The information indicating the location may be, for example, a URL (Uniform Resource Locator) or a URI (Uniform Resource Identifier).

[0483] In the above first to third embodiments and the modified examples, the compositing unit 204 combines a plurality of synthesis processes, but the compositing unit 204 may weight the synthesis processes. For example, the number of citations included in the bibliographic information of each of the plurality of synthesis processes may be used for weighting. In other words, the compositing unit 204 may specify the number of citations of each of the searched synthesis processes, and may exclude a predetermined number of synthesis processes from the synthesis processes in ascending order of the number of citations.

[0484] In the above first to third embodiments and the modified examples, the generating unit 107 associates the multiple processed words in the order of appearance in the document, but the generating unit 107 may associate the multiple processed words using a dictionary indicating the relationship between the processed words and other processed words. The generating unit 107 may also associate the multiple processed words according to a predetermined rule.

[0485] In the above first to third embodiments and the modified examples, the generated synthesis process and composite synthesis process are displayed in a graph form, but the graphed synthesis process and composite synthesis process may be corrected by the user. For example, when the user finds an error in the synthesis process displayed on the display unit 205, the user corrects the error by operating the operation device 340. In other words, when the generation unit 107 receives a signal indicating the operation result output from the operation device 340, the generation unit 107 changes the arrangement or order of multiple processing words included in the synthesis process or changes the association between the processing words and the synthesis conditions in response to the signal.

[0486] Also, the user may modify the configuration of the composite synthesis process displayed on the display unit 205, for example, by operating the operation device 340. That is, when the compounding unit 204 receives a signal indicating the operation result output from the operation device 340, it deletes or changes a plurality of processing words or material words included in the composite synthesis process in response to the signal. When such a modification of the configuration of the composite synthesis process is performed, the search unit 203 may search for a synthesis process that matches the material words and processing words included in the modified composite synthesis process. That is, a re-search of the synthesis process is performed. The compounding unit 204 generates a new composite synthesis process by compounding a plurality of synthesis processes found by the re-search, and displays the new composite synthesis process on the display unit 205. This new composite synthesis process has the same configuration (i.e., the same material words and the same processing words) as the modified composite synthesis process described above, and further includes appropriate comprehensive synthesis conditions and comprehensive characteristic values ​​according to the configuration.

[0487] In the above-mentioned first to third embodiments and the modified examples, the associations between the material words, the processing words, and the synthesis conditions in the composite synthesis flow are displayed by similar solid lines or arrows. However, the state of the solid lines or arrows showing the associations may be changed depending on the search results of the synthesis flow. For example, in the composite synthesis flow shown in FIG. 16, the number of synthesis flows in which the processing word "heat" is processed following the processing of the processing word "mix" may be greater than the number of synthesis flows in which the processing word "mill" is processed following the processing of the processing word "heat". In such a case, the compounding unit 204 may make the thickness of the arrow from the processing word "mix" to the processing word "heat" thicker than the thickness of the arrow from the processing word "heat" to the processing word "mill". This makes it possible to more clearly grasp the tendency of the synthesis process in each document.

[0488] In the above first to third embodiments and the modified examples, the bibliographic information does not include the material words of the starting material, the material words of the target material, and the characteristic values ​​of the target material, but may include them.

[0489] In the above first to third embodiments and the modified examples, the composite process generation device 100 and the composite process analysis devices 200, 240, and 250 are provided, but these devices may be included in one device or may be separate devices. For example, the composite process generation device 100 and the composite process analysis devices 200, 240, and 250 may be connected via a communication network such as the Internet. Furthermore, the multiple components that perform information processing included in these devices may be configured by one processor or circuit, or may be configured by multiple processors or circuits.

[0490] In the above first to third embodiments and the modified examples, composite process generation device 100 generates a composite process that is structured data. The generation of this composite process can be said to be a process of converting a composite process described in a document in a natural language into a composite process of structured data that can be graphed as shown in Fig. 8. Note that the composite process does not need to be described in a natural language in the document. For example, the composite process may be described in the document in a format such as a table.

[0491] In the second modification of the first embodiment, the compositing unit 204 performs clustering, which may be performed automatically or by a threshold set by the user. For example, when the compositing unit 204 clusters the pressure synthesis conditions "300 MPa", "250 MPa", "320 MPa", "450 MPa", and "460 MPa", it accepts, for example, "400 MPa" as the threshold set by the user. As a result, the compositing unit 204 clusters the set of synthesis conditions into a subset consisting of "300 MPa", "250 MPa", and "320 MPa" and a subset consisting of "450 MPa" and "460 MPa". As a result, the user can perform the desired clustering.

[0492] In each of the above embodiments and modifications, each component may be configured with dedicated hardware, or may be realized by executing a software program suitable for each component. Each component may be realized by a program execution unit such as a CPU or a processor reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory. Here, the software program that realizes the support device, generation device, and analysis device of each of the above embodiments and modifications causes a computer to execute each step included in at least one of the flowcharts shown in Figures 17A and 17B, and Figures 50A and 50B.

[0493] Furthermore, the present disclosure is not limited to the above-described embodiments and modifications, and the following cases are also included in the present disclosure.

[0494] (1) Specifically, each of the above devices is a computer system consisting of a microprocessor, ROM, RAM, hard disk unit, display unit, keyboard, mouse, etc. A computer program is stored in the RAM or hard disk unit. Each device achieves its function by the microprocessor operating according to the computer program. Here, a computer program is composed of a combination of multiple instruction codes that indicate commands for a computer to achieve a specified function.

[0495] (2) Some or all of the components constituting each of the above devices may be composed of one system LSI (Large Scale Integration). A system LSI is an ultra-multifunctional LSI manufactured by integrating multiple components on a single chip, and specifically, is a computer system composed of a microprocessor, ROM, RAM, etc. A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor operating in accordance with the computer program.

[0496] (3) Some or all of the components constituting each of the above devices may be composed of an IC card or a standalone module that can be attached to each device. The IC card or module is a computer system composed of a microprocessor, ROM, RAM, etc. The IC card or module may include the above-mentioned ultra-multifunction LSI. The IC card or module achieves its functions by the microprocessor operating according to a computer program. The IC card or module may be tamper-resistant.

[0497] (4) The present disclosure may be embodied as any of the methods described above, a computer program for implementing these methods on a computer, or a digital signal comprising a computer program.

[0498] The present disclosure may also be a computer program or a digital signal recorded on a computer-readable recording medium, such as a flexible disk, a hard disk, a CD-ROM, an MO, a DVD, a DVD-ROM, a DVD-RAM, a BD (Blu-ray (registered trademark) Disc), a semiconductor memory, etc. Also, the present disclosure may be a digital signal recorded on such a recording medium.

[0499] Furthermore, the present disclosure may involve transmitting the computer program or digital signal via a telecommunications line, a wireless or wired communication line, a network such as the Internet, data broadcasting, or the like.

[0500] The present disclosure may also provide a computer system having a microprocessor and a memory, the memory storing the above-mentioned computer program, and the microprocessor operating according to the computer program.

[0501] Furthermore, the program or digital signal may be recorded on a recording medium and transferred, or the program or digital signal may be transferred via a network or the like, so that the program or digital signal may be implemented by another independent computer system. [Industrial Applicability]

[0502] The support device, production device, and analysis device disclosed herein can appropriately support the exploration of synthesis processes, and are useful as devices or systems for material development or new material synthesis. [Explanation of symbols]

[0503] 10, 11, 12 Synthesis process search support system (support device) 13 Support equipment 100 Synthesis process generator (generator) 101 Document Selection Section 102 Document Storage Unit 103 Material word extraction unit 104 Processing Word Extraction Unit 105 Condition extraction part 106 Feature Value Extraction Unit 107 Generation part 110 First analysis device 111 Synthesis Process Extraction Unit 112 Extraction part 120 Second analysis device 121 Input section 122 Estimation Department 130 Article Database 140 Databases 141 Equipment Database 142 Materials Database 143 Project Database 144 Unit Dictionary Database 145 Customer Database 145x Organization Data Table 146 Equipment Maintenance Database 147 Sales Database 148 Estimation Results Database 200, 240, 250 Synthesis process analyzer (analytical equipment) 201 Synthesis Process Storage Unit 202 Search criteria input section 203 Search Department 204 Complex Department 205 Display section 206 Related Storage Section 207 Synthesis condition input section 208 Characteristic value estimation unit 211 Equipment Characteristics Storage Department 212 Facility Search Section 310 Computer 320 Communication Network 331 First Recording Medium 332 Second Recording Medium 340 Operation Device

Claims

1. An information processing method executed by a processor of a support device, comprising: extracting a plurality of words including a plurality of synthesis condition words indicating synthesis conditions, which are conditions for a process of generating a target material from a plurality of starting materials, by natural language processing of a sentence described in a document i (i = 1 to n); generating data associating at least one of the synthesis condition words and at least one word; The n is a natural number of 2 or more. Information processing methods.

2. The data is associated based on an appearance order or a positional relationship of the at least one synthesis condition word and the at least one word in the document i. The information processing method according to claim 1 .

3. Extracting from the document i a plurality of starting material words indicating the plurality of starting materials and a plurality of material words including a target material word indicating the target material; extracting from the document i a plurality of process words indicative of a process for producing the target material from the plurality of starting materials; By associating the plurality of material words, the plurality of processing words, and the plurality of synthesis condition words with each other, generating process information indicating a procedure for producing the target material from the plurality of starting materials; The information processing method according to claim 1 .

4. Using the plurality of synthesis condition words extracted from the document i, generate one or more comprehensive synthesis conditions comprehensively indicating a plurality of different synthesis conditions associated with a common processing word among the plurality of processing words; generating composite process information showing a procedure for producing the target material from the plurality of starting materials through the one or more comprehensive synthesis conditions; The information processing method according to claim 3.

5. In the generation of the one or more comprehensive synthesis conditions, a plurality of subsets are generated by performing clustering using the plurality of synthesis condition words extracted from the document i, and a plurality of comprehensive synthesis conditions are generated for each of the plurality of subsets. The information processing method according to claim 4.

6. Extracting a plurality of characteristic values ​​related to the target material from the document i; generating a plurality of generic characteristic values ​​indicative of the corresponding plurality of characteristic values ​​for each of the plurality of subsets; In generating the plurality of comprehensive synthesis conditions, the plurality of comprehensive synthesis conditions are associated with the plurality of comprehensive characteristic values; The information processing method according to claim 5.

7. The step of associating the plurality of material words, the plurality of processing words, and the plurality of synthesis condition words with each other includes: and associating the plurality of ingredient words with the plurality of processing words by referring to a database linking the plurality of processing words with ingredient words related to the processing words. The information processing method according to claim 3.

8. The step of associating the plurality of material words, the plurality of processing words, and the plurality of synthesis condition words with each other includes: and associating the plurality of processing words with the plurality of synthesis condition words by referring to a database in which the plurality of processing words are associated with synthesis condition words related to the processing words. The information processing method according to claim 3.

9. A program for causing a computer to execute the information processing method according to claim 1.

Images