Method for searching forms, system for searching forms, information processing device, and program

Abstract

Provided is a method for determining the degree of importance of ledgers in accordance with a query category, and changing the display order of the ledgers on the basis of the determination. Provided is a method for searching ledgers recorded in a database on the basis of document description, the method including: determining the type of each ledger; determining a category of a query that employs the description, to calculate the degree of importance of the ledgers with respect to the description; using the query to search the ledgers, by searching for the description from among sentences written in the ledgers; calculating a degree of similarity between the description employed in the query and the sentences written in the searched ledgers; and determining the display order of the ledgers from the degree of importance of each ledger and the degree of similarity of the description.

Description

Method for searching forms, system for searching forms, and information processing device

[0001] The present disclosure relates to a method for searching for a form, a system for searching for a form, and an information processing device.

[0002] A method for executing a search targeting text has been developed. Patent Document 1 describes inputting a query and classifying the query string into categories. Information having category information corresponding to the classified category is extracted as a search target, and a search process based on the query string is executed using the extracted information as the search target.

[0003] Japanese Patent Application Laid-Open No. 2006-227823

[0004] However, Patent Document 1 does not take into consideration that different forms are considered important depending on the query category. Furthermore, it does not take into consideration changing the display order of forms depending on the importance. Therefore, an object of the present disclosure is to provide a method for determining the importance of forms based on the query category and changing the display order of forms based on the determination.

[0005] The method for searching for forms disclosed herein is a method for searching for forms registered in a database based on a description in a document, which determines the type of the form, determines the category of a query using the description, and calculates the importance of the form relative to the description, searches for the form using the query by searching for the description among sentences written in the form, calculates the similarity between the sentences written in the searched form and the description used in the query, and determines the display order of the forms based on the importance of the form and the similarity of the descriptions.

[0006] The system for searching forms disclosed herein is a system for searching forms registered in a database based on a description in a document, and includes: a form importance calculation means for determining the type of the form, determining the category of a query using the description, and calculating the importance of the form relative to the description; a search means for searching the form using the query by searching for the description among sentences written in the form; a description similarity calculation means for calculating the similarity between the sentences written in the searched form and the description used in the query; and a form display order determination means for determining the display order of the forms based on the importance of the form and the similarity of the descriptions.

[0007] The information processing device disclosed herein is an information processing device that searches for forms registered in a database based on a description of a document, and includes: a form importance calculation means that determines the type of the form, determines the category of a query using the description, and calculates the importance of the form relative to the description; a search means that searches for the form using the query by searching for the description among sentences written in the form; a description similarity calculation means that calculates the similarity between the sentences written in the searched form and the description used in the query; and a form display order determination means that determines the display order of the forms from the importance of the form and the similarity of the descriptions.

[0008] The present disclosure can provide a method for determining the importance of a form based on the category of a query and changing the display order of the forms based on that determination.

[0009] Fig. 1 is a schematic diagram of a method for searching for a form according to the present disclosure; Fig. 2 is a flowchart showing an outline of a method for searching for a form according to the present disclosure; Fig. 3 is a flowchart showing a display order determination method for searching for a form according to the present disclosure; Fig. 4 is a block diagram showing the configuration of a system and an information processing device for searching for a form according to the present disclosure.

[0010] (Scope of Application of Method for Searching for Forms According to the Embodiment) A method for searching for forms will be described with reference to FIG.

[0011] The disclosed method for searching forms is for searching important descriptions in documents such as "equipment failure reports" and "quotation approval documents" from a vast collection of forms registered in a database. As shown in the first and second rows from the left in FIG. 1, important descriptions in a document are extracted. Then, as shown in the third row from the left in FIG. 1, a search is quickly performed for sections of the vast collection of forms that have similar meanings to the important descriptions. In this way, important descriptions in a document are extracted, and a fuzzy search is performed from a vast collection of forms using the important descriptions as a query.

[0012] Sentences with a similar meaning to a query statement often exist in multiple places in multiple reports. However, the importance of a report changes depending on the query. For example, in the case of an "equipment failure report," if the query is "cause of failure," then the "failure cause analysis report" or "engineer interview result report" would be important, while if the query is "date of failure," then the "failure acceptance record" would be important. This means that even if the "cause of failure" is stated in the "failure acceptance record," the "failure acceptance record" is not important.

[0013] Therefore, as shown first from the right in Figure 1, the method for searching for forms disclosed herein calculates the importance of the form for each query and determines the display order. The display order of similar parts depends on the form type and similarity. Here, "equipment failure report" is used as an example, but the method for searching for forms whose importance changes depending on other queries can be applied. The system and information processing device disclosed herein are for executing the above method.

[0014] (Description of a Method for Searching for a Form According to an Embodiment) A method for searching for a form will be described with reference to FIGS. 2 and 3. FIG.

[0015] 2, first, the importance of the form is calculated (step S201). The type of the form is determined, and the category of the query using the description of the document is determined, and the importance of the form for the description is calculated.

[0016] The type of form refers to the type of form that supports the description of a certain document, such as "Failure Cause Analysis Report" or "Engineer Interview Result Report." One method for determining the type of form is to extract the document title and determine the form type using named entity extraction, as shown in FIG. 3. A named entity refers to the text itself. For example, text such as "Failure Cause Analysis Report" is extracted directly from the form.

[0017] The second method for determining the type of a form is to use tags embedded in the form to determine the type of form. In the database, forms are tagged with tags related to the reporter, such as engineer, supervisor, or discoverer. The type of form is determined using such tags.

[0018] The third method for determining the type of form is to use a classifier to determine the type of form from the written content. The type of form is determined using a classifier that performs machine learning on multiple forms and outputs the form type when the form to be selected is input.

[0019] The fourth method for determining the type of form is to determine the type of form using a layout analyzer, which performs machine learning on the layouts of multiple forms and outputs the form type when the form to be selected is input.

[0020] These four methods for determining the type of a form may be used alone or in combination. In this way, the type of a form is determined.

[0021] Determining the category of a query using a document description means classifying the query, as shown in Figure 3. Query classification is performed by extracting named entities from the query, or extracting transformed named entities using query similarity, and classifying the query for each named entity or transformed named entity. Extracting named entities from a query means, for example, extracting "April 1, 2022" when the failure date is input. Extracting transformed named entities using query similarity means extracting "April 1, 2022" by converting the failure date from the Japanese calendar to the Gregorian calendar when the failure date is input. This assumes that a fuzzy search will be performed. Searches can also be performed for named entities that are transformed from named entities, such as "A did B."

[0022] Also, classifying queries means classifying queries into categories such as "failure date," "repair date," "reporter," "cause of failure," etc. The importance of the form for the description is calculated from the aforementioned type of form and such query classification.

[0023] Next, as shown in Fig. 2, the form is searched (step S202). The document description is searched from the sentences written on the form, and the form is searched from the database using the query.

[0024] Next, the similarity of the descriptions is calculated (step S203). The similarity between the sentences written in the searched form and the description used in the query is calculated. Because a fuzzy search is performed, the query description and the sentences on the form do not necessarily match. Therefore, as shown in FIG. 3, the similarity between the query description and the sentences on the form is calculated by vectorizing the query description and the sentences on the form and calculating the distance, or by clustering the query description and the sentences on the form.

[0025] Finally, the display order of the forms is determined (step S204) as shown in Fig. 2. The display order of the forms found by the search is determined based on the importance of the form and the similarity of the description. As shown in Fig. 3, the display order of the forms is determined by multiplying the importance of the form by the similarity of the description.

[0026] The above method provides a method for determining the importance of a form based on the query category and changing the display order of the forms based on that determination.

[0027] (Description of a system and information processing device for searching for a form according to an embodiment) A system and information processing device for searching for a form will be described with reference to FIG.

[0028] 4 , a system 400 for searching forms includes an information processing device 401 and a database 406. The information processing device 401 for searching forms according to the present disclosure includes a form importance calculation unit 402, a search unit 403, a description similarity calculation unit 404, and a form display order determination unit 405. The database 406 stores a form 407.

[0029] Although not shown, the information processing device 401 is physically composed of at least a processor (e.g., a CPU (Central Processing Unit)) that executes a program for executing processing, and a memory that stores the program. The information processing device 401 performs the processes of a form importance calculation unit 402, a search unit 403, a description similarity calculation unit 404, and a form display order determination unit 405 by executing the program.

[0030] The form importance calculation unit 402 has the function of determining the type of a form, determining the category of a query using a description, and calculating the importance of the form for the description. The calculation of the importance of a form is as described above.

[0031] The search unit 403 has a function of searching a form using a query by searching for a description in the sentences written on the form. The search for a form is as described above.

[0032] The description similarity calculation unit 404 has a function of calculating the similarity between the sentences written in the searched form and the description used in the query. The calculation of the description similarity is as described above.

[0033] The form display order determination unit 405 has a function of determining the display order of forms based on the importance of the forms and the similarity of the descriptions. The display order of forms is determined as described above.

[0034] The document importance calculation unit 402, the search unit 403, the description similarity calculation unit 404, and the document display order determination unit 405 can be read as a document importance calculation means, a search means, a description similarity calculation means, and a document display order determination means, respectively.

[0035] The database 406 includes a large-capacity storage device. The storage device is preferably a hard disk or a solid-state drive (SSD). The database 406 may be any type of database. For example, the database 406 may be a "hierarchical," "network," or "relational" database.

[0036] A huge amount of the forms 407 are stored in the storage device of the database. In particular, it is preferable that the forms 407 are tagged with meta-information or the like.

[0037] Here, the information processing device 401 and the database 406 are described separately, but the form 407 may be registered in the memory of the information processing device, and the information processing device 401 and the database 406 may be integrated. Also, some or all of the functions of the information processing device 401 may be distributed and executed by a cloud server.

[0038] Furthermore, part or all of the processing in the information processing device 401 described above can be realized as a computer program. Such a program can be stored using various types of non-transitory computer-readable media and supplied to a computer. Non-transitory computer-readable media include various types of tangible recording media. Examples of non-transitory computer-readable media include magnetic recording media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). The program may also be supplied to a computer by various types of temporary computer-readable media. Examples of temporary computer-readable media include electrical signals, optical signals, and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire or an optical fiber, or via a wireless communication path.

[0039] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0040] Each drawing is merely an example for describing one or more embodiments. Each drawing may not relate to only one particular embodiment, but may also relate to one or more other embodiments. As will be understood by those skilled in the art, various features or steps described with reference to any one drawing can be combined with features or steps shown in one or more other drawings to create, for example, an embodiment not explicitly shown or described. Not all features or steps shown in any one drawing are necessary to describe an exemplary embodiment, and some features or steps may be omitted. The order of steps described in any drawing may be changed as appropriate.

[0041] Some or all of the above embodiments can be described as in the following supplementary notes, but are not limited to the following: (Supplementary Note 1) A method for searching for forms registered in a database based on a description in a document, comprising: determining a type of the form, determining a category of a query using the description, and calculating the importance of the form with respect to the description; searching for the form using the query by searching for the description among sentences written in the form; calculating a similarity between the sentences written in the searched form and the description used in the query; and determining a display order of the form based on the importance of the form and the similarity of the description. (Supplementary Note 2) The method for searching forms according to Supplementary Note 1, in which the determination of the type of form is performed by any one of the following, or a combination thereof: extracting a document title by named entity extraction and determining the type of form, or determining the type of form using tags embedded in the form, or determining the type of form from text written on the form using a machine learning classifier, or determining the type of form from the layout of the form using a machine learning layout analyzer. (Supplementary Note 3) The method for searching forms according to Supplementary Note 1, in which the determination of the category of the query is performed by extracting named entities of the query, or extracting named entities transformed using the similarity of the query, and classifying the query for each named entity or each transformed named entity. (Supplementary Note 4) The method for searching forms according to Supplementary Note 1, in which the similarity of the description is determined by: vectorizing the text written on the form and the description and calculating the distance, or using clustering.(Appendix 5) A system for searching for forms registered in a database based on a description in a document, comprising: a form importance calculation means for determining the type of the form, determining the category of a query using the description, and calculating the importance of the form relative to the description; a search means for searching for the form using the query by searching for the description among sentences written in the form; a description similarity calculation means for calculating the similarity between the sentences written in the searched form and the description used in the query; and a form display order determination means for determining the display order of the forms based on the importance of the form and the similarity of the descriptions. (Supplementary Note 6) The system for searching for forms according to Supplementary Note 5, wherein the determination of the type of form is performed by any one of or a combination of: extracting a document title by named entity extraction and determining the type of the form, or determining the type of the form using tags embedded in the form, or determining the type of the form from text written on the form using a machine learning classifier, or determining the type of the form from the layout of the form using a machine learning layout analyzer. (Supplementary Note 7) The system for searching for forms according to Supplementary Note 5, wherein the determination of the category of the query is performed by: extracting named entities of the query, or extracting named entities transformed using the similarity of the query, and classifying the query for each named entity or each transformed named entity. (Supplementary Note 8) The system for searching for forms according to Supplementary Note 5, wherein the similarity of the description is determined by: vectorizing the text written on the form and the description and calculating the distance, or using clustering.(Supplementary Note 9) An information processing device that searches for forms registered in a database based on a description of a document, comprising: a form importance calculation means that determines the type of the form, determines the category of a query using the description, and calculates the importance of the form relative to the description; a search means that searches for the form using the query by searching for the description among sentences written in the form; a description similarity calculation means that calculates the similarity between the sentences written in the searched form and the description used in the query; and a form display order determination means that determines the display order of the forms from the importance of the form and the similarity of the descriptions. (Supplementary Note 10) The information processing device according to Supplementary Note 9, wherein the determination of the type of the form is performed by any one of or a combination of: extracting a document title by named entity extraction and determining the type of the form, or determining the type of the form using tags embedded in the form, or determining the type of the form from text written on the form by a machine learning classifier, or determining the type of the form from the layout of the form by a machine learning layout analyzer. (Supplementary Note 11) The information processing device according to Supplementary Note 9, wherein the determination of the category of the query is performed by extracting named entities of the query, or extracting named entities transformed using similarities of the query, and classifying the query for each named entity or each transformed named entity. (Supplementary Note 12) The information processing device according to Supplementary Note 9, wherein the similarity of the description is determined by: vectorizing the text written on the form and the description and calculating the distance, or using clustering.

[0042] Some or all of the elements (e.g., configurations and functions) described in Supplementary Notes 2 to 4 that are dependent on Supplementary Note 1 (e.g., method) may also be dependent on Supplementary Note 5 (e.g., system) and Supplementary Note 9 (e.g., information processing device) in the same dependency relationship as Supplementary Note 2 to Supplementary Note 4. Some or all of the elements described in any Supplementary Note may be applied to various hardware, software, recording means for recording software, systems, and methods.

[0043] This application claims priority based on Japanese Patent Application No. 2023-14930, filed February 2, 2023, the disclosure of which is incorporated herein by reference in its entirety.

[0044] 400: System for searching forms; 401: Information processing device; 402: Form importance calculation unit; 403: Search unit; 404: Description similarity calculation unit; 405: Form display order determination unit; 406: Database; 407: Form

Claims

1. A method for searching for a form registered in a database based on a description of a document, comprising: determining a type of the form, determining a category of a query using the description, and calculating the importance of the form with respect to the description; searching the form using the query by searching for the description in sentences written on the form; Calculating the similarity between the sentences written in the searched form and the description used in the query; A method for searching for forms, which determines the display order of the forms based on the importance of the forms and the similarity of the descriptions.

2. The determination of the type of the document is Extracting document titles using named entity extraction and determining the type of the form; or determining the type of the form using a tag embedded in the form; or Determining the type of the form from the text written on the form using a machine-learned classifier; or The method for searching for a form according to claim 1 , wherein the method determines the type of the form from the layout of the form using a layout analyzer that has undergone machine learning, or a combination thereof.

3. Determining the category of the query includes: Extracting a named entity of the query, or extracting a modified named entity using similarity of the query; The method for searching a form according to claim 1 , wherein the method is performed by classifying the query for each named entity or each transformed named entity.

4. The similarity of the description is The sentence written on the form and the description are vectorized and the distance is calculated, or The method for searching forms according to claim 1, wherein the method is carried out by utilizing clustering.

5. A system for searching for forms registered in a database based on document descriptions, a form importance calculation means for determining a type of the form, determining a category of a query using the description, and calculating the importance of the form with respect to the description; a search means for searching the form using the query by searching for the description among sentences written on the form; a description similarity calculation means for calculating a similarity between a sentence written on the searched form and the description used in the query; A system for searching for forms, comprising: form display order determination means for determining the display order of the forms based on the importance of the forms and the similarity of the descriptions.

6. The determination of the type of the document is Extracting document titles using named entity extraction and determining the type of the form; or determining the type of the form using a tag embedded in the form; or Determining the type of the form from the text written on the form using a machine-learned classifier; or The system for searching for forms according to claim 5, wherein the type of the form is determined from the layout of the form by a layout analyzer that has undergone machine learning, or a combination thereof.

7. Determining the category of the query includes: Extracting a named entity of the query, or extracting a modified named entity using similarity of the query; The system for searching forms according to claim 5 , wherein the system classifies the queries by the named entities or the transformed named entities.

8. The similarity of the description is The sentence written on the form and the description are vectorized and the distance is calculated, or 6. The system for searching forms according to claim 5, wherein the system uses clustering.

9. An information processing device that searches for forms registered in a database based on a description of a document, a form importance calculation means for determining a type of the form, determining a category of a query using the description, and calculating the importance of the form with respect to the description; a search means for searching the form using the query by searching for the description among sentences written on the form; a description similarity calculation means for calculating a similarity between a sentence written on the searched form and the description used in the query; an information processing device comprising: a form display order determination means for determining a display order of the forms based on the importance of the forms and the similarity of the descriptions;

10. A method for searching for a form registered in a database based on a description of a document, comprising: determining a type of the form, determining a category of a query using the description, and calculating the importance of the form with respect to the description; searching the form using the query by searching for the description in sentences written on the form; Calculating the similarity between the sentences written in the searched form and the description used in the query; A program that causes an information processing device to determine the display order of the forms based on the importance of the forms and the similarity of the descriptions, and to search for the forms.