Document analysis device, document analysis method, and document analysis program

Abstract

It assists in creating regulatory documents that consist of multiple documents. [Solution] A document analysis device 1 comprising: a receiving means for receiving input of a prescribed document composed of multiple divided documents as data to be analyzed; a storage unit 1A that stores requirements applicable to the divided documents; an analysis unit 11 that determines whether the requirements are met for each of the multiple divided documents; a first output means that outputs whether the requirements are met for each of the multiple divided documents as an analysis result of the prescribed document; an identification unit 12 that identifies a divided document to be modified to satisfy any requirement that is not met in any of the divided documents based on the analysis results; and a second output means that outputs information of the divided document identified as the target for modification for each requirement.

Description

Technical Field

[0001] The present invention relates to a document analysis apparatus, a document analysis method, and a document analysis program.

Background Art

[0002] Conventionally, in the creation of various documents, especially documents with expertise, systems for assisting in creation have been proposed (see Patent Documents 1 to 3).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0004] In this regard, so-called security documents (prescribed documents) are required to conform to various guidelines and are composed of multiple documents. Therefore, the work of checking and correcting documents requires a great deal of time, and since the human resources capable of creating such documents are limited, the human burden and cost burden are not small.

[0005] Therefore, an object of the present invention is to assist in the creation of prescribed documents composed of multiple documents.

Means for Solving the Problems

[0006] To achieve the above objective, a document analysis device according to one aspect of the present invention includes: a receiving means for receiving input of a prescribed document composed of multiple divided documents as data to be analyzed; a storage means for storing requirements applicable to the divided documents; an analysis means for determining whether the requirements are satisfied for each of the multiple divided documents; a first output means for outputting whether the requirements are satisfied for each of the multiple divided documents as an analysis result of the prescribed document; and, based on the analysis results, if there are requirements that are not satisfied in any of the divided documents, an identification means for identifying a divided document to be modified to satisfy those requirements. The system includes a second output means that outputs information on the segmented documents identified as targets for modification for each of the aforementioned requirements.

[0007] Based on the analysis results, if there are any divided documents that do not fully satisfy certain requirements, the aforementioned identification means may identify the divided document with the highest percentage of satisfaction of those requirements as the divided document to be modified to satisfy those requirements.

[0008] Based on the analysis results, if there are any predetermined requirements that are not satisfied in any of the divided documents, the aforementioned identification means may identify a divided document that is highly similar to the predetermined requirements as a divided document to be modified to satisfy those predetermined requirements.

[0009] The aforementioned identification means may, based on the analysis results, identify the split document selected by the user as the split document to be modified in order to satisfy the predetermined requirements if any of the split documents are not satisfied.

[0010] The above requirements may be associated with comments made during modification, and the second output means may output information on the segmented document identified as the target of modification and the comments for each of the above requirements.

[0011] The system further comprises an AI control means for transmitting instructions to an artificial intelligence unit, and an acquisition means for acquiring a modified divided document from the artificial intelligence unit, wherein the AI ​​control means transmits to the artificial intelligence unit, along with a modified instruction to modify the divided document identified as the target for modification, and an output instruction to output the modified divided document.

[0012] A document analysis method according to another aspect of the present invention involves a computer having a storage means that stores requirements applicable to divided documents, performing the following steps: an input process that receives a prescribed document composed of a plurality of divided documents as data to be analyzed; an analysis process that analyzes whether the requirements are satisfied for each of the plurality of divided documents; a first output process that outputs whether the requirements are satisfied for each of the plurality of divided documents as an analysis result of the prescribed document; an identification process that identifies a divided document to be modified to satisfy a requirement if there is a requirement that is not satisfied in any of the divided documents based on the analysis result; and a second output process that outputs information of the divided document identified as the target for modification for each requirement.

[0013] A document analysis program according to another aspect of the present invention causes a computer having a storage means that stores requirements applicable to a divided document to perform the following: an acceptance process that accepts a prescribed document composed of a plurality of divided documents as data to be analyzed; an analysis process that analyzes whether the requirements are satisfied for each of the plurality of divided documents; a first output process that outputs whether the requirements are satisfied for each of the plurality of divided documents as the analysis result of the prescribed document; an identification process that identifies a divided document to be modified to satisfy a requirement if there is a requirement that is not satisfied in any of the divided documents based on the analysis result; and a second output process that outputs information of the divided document identified as the target for modification for each requirement.

[0014] Computer programs can be provided by storing them on various data-readable storage media, or by making them available for download via networks such as the Internet. [Effects of the Invention]

[0015] According to the present invention, the burden of creating a regulatory document composed of a plurality of documents is reduced.

Brief Description of the Drawings

[0016] [Figure 1] It is a functional block diagram showing the functional configuration of a document analysis device according to an embodiment of the present invention. [Figure 2] It is a process flow diagram showing the flow of processing executed by the document analysis device. [Figure 3] It is a process flow diagram showing the flow of analysis processing executed by the document analysis device. [Figure 4] It is an example of a screen output by the document analysis device. [Figure 5] It is an example of a screen output by the document analysis device. [Figure 6] It is an example of a screen output by the document analysis device. [Figure 7] It is an example of a screen output by the document analysis device.

Embodiments for Carrying Out the Invention

[0017] ●Overview Hereinafter, a document analysis device 1 according to a first embodiment of the present invention will be described with reference to the drawings. The document analysis device 1 is a device that analyzes a created document and supports the creation of an appropriate document. The document analysis device 1 analyzes, in particular, a security regulatory document as data to be analyzed and supports its creation. A security regulatory document usually consists of a plurality of documents, and fulfillment of requirements indicated by various guidelines and the like is required. Hereinafter, such a plurality of documents will be referred to as divided documents, and a document composed of a plurality of divided documents will be referred to as a regulatory document. However, regardless of this embodiment, the present invention can be applied to documents in other fields other than security regulatory documents as long as they are documents composed of a plurality of documents and for which fulfillment of predetermined requirements is required.

[0018] ●Functional configuration As shown in FIG. 1, the document analysis device 1 is configured to be communicable with the user terminal 2 via a network NW such as the Internet. However, the document analysis device 1 may be composed of a plurality of hardware configurations. Also, a part or all of the functions of the document analysis device 1 may be implemented in the user terminal 2 as an application program.

[0019] The user terminal 2 is a terminal used by a user who uses the service provided by the function of the document analysis device 1. The user terminal 2 is realized by various terminals capable of executing a computer program, such as a portable terminal or a tablet terminal capable of transmitting and receiving data, as well as a so-called personal computer. Also, the users include not only those who receive assistance in creating documents, but also those who provide information necessary for executing the functions of the document analysis device 1 or manage the execution of such functions.

[0020] This user terminal 2 includes at least communication processing means such as a browser program for executing data transmission and reception with the document analysis device 1, and input / output means such as a touch display for inputting and outputting data. Also, the user terminal 2 can create, edit, and view a prescribed document or a divided document, or view information created by the document analysis device 1 via an application program for creating, editing, and viewing documents. Note that the application for creating, editing, and viewing a prescribed document or a divided document may be a dedicated application, or the above-described processing may be enabled via a general-purpose web browser. Also, the functions provided by the document analysis device 1 may be incorporated as an add-in into a known application program for creating, editing, and viewing documents.

[0021] The document analysis device 1 is implemented by, for example, a server equipped with a processing unit such as a CPU (Central Processing Unit), a computer program executed by the CPU, and internal memory such as RAM (Random Access Memory) and ROM (Read Only Memory). As a result, the document analysis device 1 constitutes a functional block consisting of an information storage unit 1A, an analysis unit 11, a identification unit 12, an extraction unit 13, and a communication processing unit 14, as shown in Figure 1.

[0022] <Information storage section 1A> The information storage unit 1A stores various types of information. For example, requirements data and comment data are stored in this information storage unit 1A. Requirements data consists of multiple requirements that apply to the segmented documents that make up the regulatory document being analyzed. Requirements data may be registered by the user. Furthermore, requirements data may be a structured set of guidelines or keywords specified by the user. The requirements data may be grouped by similar purposes or meanings.

[0023] Comment data consists of comments attached to the split documents as a result of the analysis process. The comments include a wide range of information that helps users fulfill the requirements, such as suggestions for the direction of revision, example sentences that should be specifically included in the split documents, interview items and reasons for interviewing when creating or revising the split documents, risks that may arise from not meeting the requirements, and solutions to those risks.

[0024] This comment data is associated with requirements, and if a split document does not satisfy a predetermined requirement, a comment corresponding to that predetermined requirement will be attached as a comment to the split document when presenting the analysis results to the user. Furthermore, comments may be attached to the segmented documents identified as requiring modification as a result of the analysis, or they may be attached to all segmented documents that have been analyzed. Furthermore, multiple comments may be associated with a single requirement, and comments may be attached in a manner that reflects how the requirement is met, or in relation to the content of the divided document to which the requirement applies.

[0025] The information storage unit 1A may also store the standard documents that are subject to analysis. While there are no particular restrictions on the data format of the standard documents, structured JSON or Markdown formats are preferred.

[0026] <Analysis Department 11> The analysis unit 11 determines whether the requirements are met for each of the multiple divided documents. That is, it performs a one-to-one analysis of all the multiple divided documents and their requirements. This allows it to determine the status of requirement fulfillment in each divided document, i.e., which requirements are met in which divided document, and which are not met in any of the divided documents.

[0027] <Specific part 12> Based on the analysis results by the analysis unit 11, the identification unit 12 identifies the divisional document to be modified to satisfy any requirement that is not met in any of the divisional documents. Based on the analysis results by the analysis unit 11, if there are any predetermined requirements that are not fully satisfied in any of the divided documents, the identification unit 12 may identify a predetermined divided document as a divided document to be modified to satisfy those predetermined requirements.

[0028] In this case, the specified segmented document may be the one with the highest percentage of fulfillment of the given requirements in the first example, or the one with the lowest percentage of fulfillment of the given requirements. The percentage of fulfillment can be determined, for example, by the fulfillment rate of the keywords set as requirements. Furthermore, in the second example, the identified segmented document may be one that has a higher degree of similarity to the given requirements than others. Similarity can be determined, for example, by the context between the requirements and the segmented document, and the correlation between the words and phrases used in each. Furthermore, in the third example, the split document to be identified may be one selected by the user. For example, the analysis results could show the user whether all split documents meet the requirements, and then identify the split document selected by the user as the one to be modified.

[0029] Notwithstanding this example, it is also possible to designate multiple split documents as targets for modification and set modification priorities without uniquely identifying the split documents to be modified. For example, when identifying documents to be modified according to the percentage of satisfaction of a predetermined requirement, it is possible to designate split documents with a satisfaction percentage below a certain level as targets for modification, and then set modification priorities in order from those with the lowest satisfaction percentages, or in order of similarity to the requirement. Alternatively, when identifying documents to be modified according to their similarity to a predetermined requirement, it is possible to designate split documents with a similarity above a certain level as targets for modification, and then set modification priorities in order from those with the highest similarity, or in order of the degree to which the requirement is satisfied.

[0030] Furthermore, if there are divided documents that satisfy the prescribed requirements and divided documents that partially do not satisfy the prescribed requirements, the latter may be designated as the target for correction. This makes it possible to detect and correct inconsistencies even when there are inconsistencies in multiple regulation documents. For example, when revising a regulation document, if the text relating to the prescribed requirements spans multiple divided documents, all of those divided documents would normally be subject to correction. However, by identifying the divided document that partially does not satisfy the prescribed requirements and designating it as the target for correction, it is possible to prevent corrections from being missed.

[0031] Furthermore, if the requirements data is grouped by similar purpose or meaning, the split documents to be modified may be identified based on the satisfaction rate of each group. For example, for multiple requirements belonging to the same group, the satisfaction rate in one split document can be calculated individually and then summed up to determine the satisfaction rate of the grouped requirements. Then, the group with the highest or lowest satisfaction rate can be identified as the one to be modified. Similarly, when identifying based on similarity, for multiple requirements belonging to the same group, the similarity in one split document can be calculated individually and then summed up to determine the similarity of the grouped requirements, and the group with the highest similarity can be identified as the one to be modified.

[0032] <Extract part 13> Based on the analysis results from the analysis unit 11, the extraction unit 13 extracts comments associated with any unfulfilled requirements from the information storage unit 1A. The extracted comments are then output, for example, attached to the segmented document to be modified, as identified by the identification unit 12. In this example, the state in which the prescribed requirements are not met includes not only the state in which the prescribed requirements are not met at all, but also the state in which the requirements are not fully met, that is, the state in which some of the prescribed requirements are met, but some parts are not. Furthermore, the extraction unit 13 may extract comments from all segmented documents for which the requirements have been determined to be met, regardless of whether the requirements are met or not.

[0033] <Communication Processing Unit 14> The communication processing unit 14 performs data transmission and reception processing with the user terminal 2 via a network NW such as the Internet, in accordance with a predetermined communication protocol. In this example, the communication processing unit 14 constitutes an input / output means that receives input of the specified document to be analyzed from the user terminal 2 connected via the network NW and outputs the analysis results. With the communication processing unit 14 as an input / output means, the document analysis device 1 outputs (transmits) to the user terminal 2 whether predetermined requirements are met for each of the multiple divided documents that constitute the specified document, as an analysis result of the specified document. The document analysis device 1 also constitutes a second output means that outputs information of the divided documents identified as targets for modification for each requirement. In this case, the communication processing unit 14 also outputs (transmits) comments on the identified divided documents to the user terminal 2.

[0034] ●Processing flow and screen examples The following describes the processing flow diagrams shown in Figures 2 and 3, and the screen examples shown in Figures 4 through 8. Figure 3 shows the overall processing flow, while Figures 3(a) and 3(b) show detailed examples of the analysis process flow. Prior to the analysis process, the user sets the requirements to be used for the analysis of the specified document in advance (S11). Figure 4 shows an example of a requirements setting screen displayed on user terminal 2. In this example screen, the requirements data called "Web Assessment," which was registered in information storage unit 1A, is retrieved according to the user's specifications. On the screen, multiple requirements that make up the requirements data are displayed in a hierarchical structure. Specifically, in the example shown, under the item "1. Organization," there is a requirement "1.1 Define the roles and responsibilities of those who handle the information system" (hereinafter referred to as "First Requirement"), and under this requirement, there are three further requirements such as "1.1.1 Define the roles and responsibilities of users, developers, operators, applicants, approvers, and executors" (hereinafter referred to as "Second Requirement" without distinction from each other).

[0035] The first requirement has an "Automatic Analysis" button, which, when activated, sets all sub-requirements under the first requirement to be applied during the analysis process. Similarly, each sub-requirement also has an activation button, which, when activated, allows you to choose whether or not to apply the target requirement during the analysis process. Furthermore, requirement data can be registered in the information storage unit 1A in advance, and the user can retrieve the appropriate requirement data during the analysis process and set whether or not to apply it during the analysis process.

[0036] The document analysis device 1 receives input of the specified document to be analyzed from the user terminal 2 (S12). Figure 5 shows an example of a screen for confirming the execution details of the analysis process. On the screen, the data name of the regulation document to be analyzed (shown as "Sample Internal Regulations_ver1" in the figure) is displayed in the "Select Document to Use" area. In addition, the data name of the requirement data to be applied in the analysis process (shown as the guideline name "Web Assessment" and version "1.0.5" in the figure) is displayed in the "Assessment Settings" area. Furthermore, an assessment name can be set separately for each assessment to be executed, and in the example shown, the name of the assessment to be executed is displayed as "Internal Security Document Assessment". In the upper right of the diagram, there is an execution button to request the execution of the analysis process. The user confirms the details of the analysis process on the confirmation screen, and if there are no problems, presses the execution button to request the document analysis device 1 to execute the analysis process.

[0037] The document analysis device 1 performs the analysis process of the specified document to be analyzed in response to an execution request from the user (S13). In the analysis process, as shown in Figure 3(a) in one example, user-specified requirement data is applied to the specification document to be analyzed. First, the analysis unit 11 determines whether the requirements are met for each segmented document that makes up the specification document (S131). Subsequently, the extraction unit 13 extracts comments associated with the applied requirements from the information storage unit 1A for multiple segmented documents (S132).

[0038] In another example of the analysis process, as shown in Figure 3(b), in step S131, after determining whether the requirements are met for each divided document (S131), the identification unit 12 identifies the divided document to be modified based on whether the requirements for each divided document are met (S133). If there are any requirements that are not met by any of the divided documents, the extraction unit 13 extracts comments associated with those unmet requirements from the information storage unit 1A (S134).

[0039] When the document analysis device 1 completes the analysis process, it outputs the analysis results to the user terminal 2 (S14). Figure 6 shows an example of a screen displaying the analysis results. The illustrated screen simply shows whether each requirement (Requirement 1) has been met. The fulfillment of a requirement is shown in the form of a score and a status corresponding to that score. In the example shown, a five-level score (with a score of 1 being the highest and a score of 5 being the lowest) is displayed according to the percentage of the specified requirements that have been met. For example, if all of the Requirements 2 under Requirements 1 are met, a score of 1, the highest rating, is given, and if only one of the Requirements 2 under Requirements 1 is met, a score of 3 is given.

[0040] Figure 7 is an example screen showing in detail whether each requirement has been met. This example screen includes comments added to all segmented documents where the requirements have been determined to be met. On the screen, the level of compliance for each requirement is displayed in the "Judgment" column as "Compliant," "Partially Compliant," or "Non-Compliant." In addition, each segmented document has a comment in the "Reason for Judgment" column, and the relevant section within the segmented document is displayed in the "Quoted Text" column. Furthermore, when the identification unit 12 identifies the divided document to be modified, for example, only the divided document to be modified is displayed, and comments associated with the requirements that should be satisfied in that divided document are attached.

[0041] Furthermore, in the example screens shown in Figures 6 and 7, it may be possible to narrow down the display of requirements and divided documents according to scores, statuses, etc., or to display analysis results by requirement or by divided requirement.

[0042] According to the document analysis device 1 of this embodiment, the burden of creating a standard document composed of multiple documents is reduced. In particular, it is easy to understand which of the divided documents satisfies each requirement, and the divided documents that need to be modified are also presented, making it easy to understand the items and content that need to be checked and corrected.

[0043] ● Variation Notwithstanding the above embodiment, the document analysis device 1 can be modified to include or replace the following configurations. In other words, in one embodiment, the specific unit 12 may generate a proposed revision for the segmented document identified as a target for revision and present it to the user.

[0044] In realizing such an embodiment, the document analysis device 1 is configured to communicate with, for example, an artificial intelligence unit and has an AI control unit.

[0045] The AI ​​control unit is a functional unit that controls the input to the artificial intelligence unit. The AI ​​control unit generates instructions for the artificial intelligence unit. These instructions are, for example, prompts, but are not limited to those that the artificial intelligence unit can interpret. This AI control unit transmits to the artificial intelligence unit, along with the segmented document identified as the target for modification, a modification instruction to modify the segmented document, and an output instruction to output the modified segmented document.

[0046] The Artificial Intelligence (AI) unit is an artificial intelligence (AI) equipped with appropriate learning models. A learning model (also called a machine learning model) refers to a learning model based on a machine learning algorithm. Specific machine learning algorithms include nearest neighbors, naive Bayes, decision trees, and support vector machines. Another example is deep learning, which uses neural networks to generate features and connection weights for learning. The Artificial Intelligence unit can apply the above algorithms as appropriate.

[0047] Furthermore, the artificial intelligence unit may be an artificial intelligence equipped with language models such as transformers including BART (Bidirectional and Auto-regressive Transformer), BERT (Bidirectional Encoder Representations from Transformers), or GPT (Generative Pretrained Transformer, including GPT-1, GPT-2, GPT-3, and GPT-4), and especially learning models such as large language models (LLM). Using such a learning model, the artificial intelligence unit generates and outputs example sentences in response to example sentence generation instructions.

[0048] The artificial intelligence department possesses pre-trained models developed through machine learning as needed. The training data may be provided by administrators or other relevant parties, or it may include information collected from the internet or other sources.

[0049] In one example, the artificial intelligence unit, based on the correction instructions generated by the AI ​​control unit, uses a trained model to take the divided document as input and output a corrected divided document. In this case, the artificial intelligence unit has a trained model that takes the divided document as input and outputs an example sentence. The artificial intelligence unit may also take only a portion of the divided document as input, in which case it will output a divided document with that portion corrected.

[0050] In other examples, the artificial intelligence unit may be connected to a predetermined database that stores comments such as example sentences for divided documents. In this case, the AI ​​control unit may instruct the artificial intelligence unit to extract comments corresponding to the divided document to be modified, and the artificial intelligence unit may extract example sentences from the predetermined database in response. Furthermore, in the modification instructions from the AI ​​control unit, the predetermined database may be specified as a database to be searched by RAG (Retrieval-Augmented Generation). That is, the modification instructions sent from the AI ​​control unit to the artificial intelligence unit may include an instruction requesting a lookup of the predetermined database, and in response, the artificial intelligence unit may refer to the predetermined database to modify the divided document or extract comments such as example sentences.

[0051] Furthermore, the output of the corrected segmented document and the output of comments, as described above, may be performed together. That is, the artificial intelligence unit may have a learning model that takes a correction instruction for a segmented document that does not satisfy a predetermined requirement and a comment extraction instruction for that segmented document as input, and outputs the corrected segmented document and comments.

[0052] Furthermore, AI functionality may also be applied to the process by the identification unit 12 for identifying the segmented documents to be modified that satisfy the requirements. For example, the document analysis device 1 may be equipped with an AI control unit as described above, and configured to communicate with the artificial intelligence unit. The AI ​​control unit sends instructions to the artificial intelligence unit to identify the segmented documents to be modified and to output the identified results. In response, the artificial intelligence unit calculates the percentage of requirements satisfied for each segmented document and the similarity between each segmented document and the requirements, and outputs the segmented document with the highest percentage of requirements satisfied and the segmented document with the highest similarity to the requirements as the segmented documents to be modified. The artificial intelligence unit includes a learning model that takes segmented documents and requirements as input and outputs the percentage of requirements satisfied and the similarity in the segmented documents.

[0053] Furthermore, the AI ​​function may also be applied when the extraction unit 13 retrieves comments from the storage unit 1A. In this case, the artificial intelligence unit has a learning model that takes the divided document as input, along with the status of requirement fulfillment and the instruction to output comments, and outputs comments according to the status of requirement fulfillment.

[0054] Furthermore, AI functionality may also be applied to determining the status of requirement fulfillment. For example, the AI ​​control unit sends an instruction to the artificial intelligence unit to determine the status of requirement fulfillment in the divided document and an instruction to output a comment corresponding to the fulfillment status. In response, the artificial intelligence unit determines the status of requirement fulfillment and outputs a comment corresponding to that status. The artificial intelligence unit includes a first learning model that takes the divided document and requirements as input and outputs the status of requirement fulfillment in the divided document, and a second learning model that takes the fulfillment status as input and outputs a comment corresponding to that status. However, the output of the comment may be performed by referring to a predetermined database as described above. [Explanation of Symbols]

[0055] 1: Document analysis device 1A: Information storage section 11: Analysis department (analysis means) 12: Specification part (specification means) 13:Extraction part 14: Communication processing unit (output means) 2: User terminal NW: Network

Claims

1. The data to be analyzed includes a receiving means for accepting input of a standard document composed of multiple divided documents, A storage means that stores the requirements applicable to the divided document, For each of the aforementioned multiple divided documents, an analysis means for determining whether the requirements are met, As a result of analyzing the aforementioned provision document, a first output means outputs whether or not the requirements are met for each of the multiple divided documents, Based on the analysis results, if there are any predetermined requirements that are not met in any of the divided documents, the means for identifying the divided document with the highest percentage of satisfaction of said predetermined requirements is to be modified to satisfy those predetermined requirements, The system includes a second output means that outputs information of the segmented documents identified as targets for modification for each of the aforementioned requirements, Document analysis device.

2. The data to be analyzed includes a receiving means for accepting input of a standard document composed of multiple divided documents, A storage means that stores the requirements applicable to the divided document, For each of the aforementioned multiple divided documents, an analysis means for determining whether the requirements are met, As a result of analyzing the aforementioned provision document, a first output means outputs whether or not the requirements are met for each of the multiple divided documents, Based on the analysis results, if there are any predetermined requirements that are not satisfied in any of the divided documents, the following means are provided to identify a divided document that has a high degree of similarity to the predetermined requirements as a divided document to be modified to satisfy those predetermined requirements: The system includes a second output means that outputs information of the segmented documents identified as targets for modification for each of the aforementioned requirements, Document analysis device.

3. The aforementioned requirements are associated with comments made during the revision process. The second output means outputs, for each requirement, the information of the segmented document identified as the target for modification and the comment. The document analysis apparatus according to claim 1 or 2.

4. AI control means that transmits instructions to the artificial intelligence unit, The artificial intelligence unit further comprises an acquisition means for acquiring a modified segmented document, The AI ​​control means is The artificial intelligence unit receives, along with the segmented documents identified as targets for modification, a modification instruction to modify the said segmented documents, Send an output instruction to output the corrected split document. The document analysis apparatus according to claim 1 or 2.

5. A computer having storage means that stores the requirements applicable to the divided document, The data to be analyzed includes an input process that accepts a standard document composed of multiple such divided documents, For each of the aforementioned multiple divided documents, an analysis process is performed to analyze whether the requirements are met, As a result of the analysis of the aforementioned provision document, a first output process outputs whether or not the requirements are met for each of the multiple divided documents, Based on the analysis results, if any of the divided documents fail to satisfy certain requirements, the following process is performed to identify the divided document with the highest percentage of satisfaction of those requirements as the divided document to be modified to satisfy those requirements: For each of the above requirements, a second output process is performed to output information of the segmented documents identified as targets for modification. Document analysis methods.

6. A computer having a storage means that stores requirements applicable to divided documents, The data to be analyzed includes an input process that accepts a standard document composed of multiple such divided documents, For each of the aforementioned multiple divided documents, an analysis process is performed to analyze whether the requirements are met, As a result of the analysis of the aforementioned provision document, a first output process outputs whether or not the requirements are met for each of the multiple divided documents, Based on the analysis results, if any of the divided documents fail to satisfy certain requirements, the following process is performed to identify the divided document with the highest percentage of satisfaction of those requirements as the divided document to be modified to satisfy those requirements: For each of the above requirements, a second output process is performed to output information of the segmented documents identified as targets for modification. Document analysis program.

7. A computer having storage means that stores the requirements applicable to the divided document, The data to be analyzed includes an input process that accepts a standard document composed of multiple such divided documents, For each of the aforementioned multiple divided documents, an analysis process is performed to analyze whether the requirements are met, As a result of the analysis of the aforementioned provision document, a first output process outputs whether or not the requirements are met for each of the multiple divided documents, Based on the analysis results, if any of the divided documents fail to satisfy a predetermined requirement, the following process is performed to identify the divided document with a high degree of similarity to that requirement as the divided document to be modified to satisfy that requirement: For each of the above requirements, a second output process is performed to output information of the segmented documents identified as targets for modification. Document analysis methods.

8. A computer having a storage means that stores requirements applicable to divided documents, The data to be analyzed includes an input process that accepts a standard document composed of multiple such divided documents, For each of the aforementioned multiple divided documents, an analysis process is performed to analyze whether the requirements are met, As a result of the analysis of the aforementioned provision document, a first output process outputs whether or not the requirements are met for each of the multiple divided documents, Based on the analysis results, if any of the divided documents fail to satisfy a predetermined requirement, the following process is performed to identify the divided document with a high degree of similarity to that requirement as the divided document to be modified to satisfy that requirement: For each of the above requirements, a second output process is performed to output information of the segmented documents identified as targets for modification. Document analysis program.

Images