JP2009245232A - Dedicated rule editor for generating rule definition of problem extraction from free description sentence of medical observation document - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a means for allowing a user to update a rule in an easy operation without requiring the user to learn a definition of the rule which is to be flexibly and continuously updated for language analysis, in a method for performing diagnosis support by performing the language analysis of a free description sentence of a medical observation document of a report system or an electronic medical chart and extracting a clinical problem or a disease name, the method being installed to a medical information support system. <P>SOLUTION: This dedicated rule editor has: a display part displaying a display face showing an idea of a person who updates the rule as it is when necessary; an input part for performing an operation input of a corresponding part in the display face; a keyword storage part and a rule storage part storing a keyword on the display face and the problem respectively; an inspection part verifying that the stored rule or keyword is consistent; and a conversion part converting the stored rule into the rule. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention is a method for supporting medical practice by extracting a problem related to a disease name, symptom, examination, treatment, treatment, etc. according to a rule for language analysis of a free description sentence of a medical finding document in an electronic medical record or a medical finding report system. The present invention relates to a dedicated rule editor for generating a rule definition for extracting a problem through language analysis only by an input editing operation of a free description sentence. This dedicated rule editor is suitable as a rule generation support tool for use in a “problem extraction engine” incorporated in a medical risk warning system, a clinician decision support system, etc. by supporting and monitoring a medical practice.

  In recent years, medical information such as medical records, medication calendars, and instruction screens has been accumulated as electronic information in medical institutions such as hospitals and clinics as electronic medical record systems and medical finding report systems have become widespread. These electronic medical information can be freely set and processed for each department, doctor, disease, initial visit / revisit, etc., and the medical information can be varied by multiple people anytime and anywhere. Since it can be used, it is effectively used for the operation of medical institutions and the support of medical personnel such as improving the efficiency of medical practices and improving services.

  However, the utilization method of the medical information obtained from the electronic medical record system and the medical observation report system is limited to such a utilization method so far. The inventors of the present invention have intensively studied on the advanced utilization method based on the idea that useful medical information obtained from the electronic medical chart system can be used at a higher level. As a result, we have developed and proposed a method for extracting problems related to disease names, symptoms, examinations, treatments, treatments, etc. by analyzing and evaluating medical information obtained from medical findings using natural language understanding technology. The present invention makes it easier to use the problem extraction method from the free description sentence of the medical finding document and applies the automatic learning technology of language knowledge developed by the inventors for practical use, thereby analyzing and evaluating medical information This is to develop a dedicated rule editor that learns information at high speed and generates rule definitions for problem extraction.

  Many electronic medical records in the electronic medical record system and report documents in the medical observation report system have formats that are specified in order to reduce their input, such as columns that can be selected from the choices and columns that are described as numerical values. In most cases, a manual description column is provided. These formal description fields are effectively used in the medical information system as easy-to-handle data for the operation of hospital organizations and the support of medical personnel.

  At the same time, the electronic medical record and medical findings report are also provided with a section where doctors can freely write down as notes, such as findings and discharge summary. These free descriptions such as findings and discharge summaries often contain important items for diagnosis, such as matters to be noted, progress matters, and grounds for diagnosis, and contain many important items that are useful for problem extraction. Yes. However, since the medical finding document is freely described in an atypical sentence, it is difficult to handle it as it is, and in order to extract a problem, it is necessary to perform a language analysis using a natural language processing technique for the free description sentence.

  The research and development team, including the inventor, conducts research on a medical risk warning system based on a high-speed learning technique for medical diagnosis knowledge, and the premise of the present invention is "extraction of problems from free sentences using natural language understanding technology" The technology was developed and presented at the 27th Medical Information Union Conference. The gist of this paper is to propose a method of extracting a problem by defining a rule of analysis when a linguistic analysis is performed on a doctor's findings (Non-patent Document 1).

Junichi Aoe, Kumiko Kiyoi, Yoshihiro Yamakawa, Yasushi Matsumura, Sachiko Yoshimoto, Tetsuya Yamaguchi, Mai Yoshida, “Extracting Problem from Free Text Using Natural Language Understanding Technology”, 27th Medical Information Conference, 10504

As described above, in response to a request for diagnosis support using medical information in a free description sentence of a medical finding document, a method of supporting the problem by extracting a problem according to a rule for language analysis of the free description sentence (Non-patent Document 1) In this proposal, it is necessary to define a rule in order to extract a problem by linguistically analyzing a free description sentence of a medical finding document. However, in order for a user to define a rule in this way, there is a problem that it is necessary to consider a logical structure for the definition of the rule and to learn and master the definition method.
It is practically impossible and inappropriate to make a doctor who is a user of an electronic medical record system or a medical observation report system overcome this problem. For this reason, it has been a big obstacle when using the problem extraction method from the free description of the medical findings document developed.

  In order to solve such problems, the present inventor has sought to develop a tool (dedicated rule editor) that can generate language analysis rules by a simple free description input editing operation.

  In the development, when the user performs an input operation considering the case of medical treatment from the description in the medical observation document, the input is performed from a method that has been conceived, so the tool may have a function to input an idea. It is desired. For this reason, it is necessary to generate rules for language analysis from the tool.

Accordingly, the present invention relates to a method for extracting a problem based on a rule for language analysis of a free description sentence of a medical finding document. Has succeeded in realizing a dedicated rule editor that generates the text only by the free text input operation.

  The present invention has the following configuration as means for solving the above technical problem.

The first invention to be patented is a method for extracting a medical problem such as disease name, symptom, examination, treatment, and treatment by defining a rule for language analysis of a free description sentence of a medical finding document. Is a dedicated rule editor that generates rules by inputting the definition of rules for extracting problems, and it is a keyword, a category that is a keyword set that summarizes keyword fluctuations and synonyms, a problem, and whether it is a problem An input unit for performing an input editing operation on a rule that is a combination of a category and a keyword for diagnosing
A display section for displaying the input and edited problems, rules, categories, and keywords;
A keyword storage unit connected to the input unit or the display unit and storing the input keyword;
A rule storage unit that is connected to the input unit or the display unit and stores an input category and rule,
An inspection unit that is connected to the keyword storage unit and the rule storage unit, and verifies whether there is a contradiction in the combination of circulation and conflicting rules,
A dedicated rule editor comprising a conversion unit that is connected to the keyword storage unit and the rule storage unit and generates a stored keyword and rule as a rule for language analysis.

The first invention is a basic invention of a dedicated rule editor that generates a problem extraction rule definition from a free description sentence of a medical finding document. The dedicated rule editor includes the input unit, the display unit, the keyword storage unit, the rule storage unit, the inspection unit, and the conversion unit as described above, and the input unit, the display unit, the inspection unit, and the conversion unit are all included. The keyword storage unit and the rule storage unit are connected. The characteristics of the constituent elements are as follows.

  The display section of the dedicated rule editor described in the first invention has a tabular display surface, and in each row of the table, a cell for inputting and displaying a problem name, a keyword and a category for estimating the problem are input. The cell to be displayed, the cell to input and display the category name, the keywords to be included in the category, and the plurality of cells to input and display the category are provided as basic formats. A cell for selecting whether to extract a condition, a problem name, a plurality of keywords as a condition, and a plurality of cells defining a plurality of categories are arranged. This is a table having a configuration in which a plurality of cells defining a plurality of keywords and a plurality of categories are arranged. This is a feature of the second invention to be patented.

The second invention is an invention that is specifically specified for the configuration of the display unit of the dedicated rule editor according to the present invention. That is, the display unit can display the name of the problem being defined, the category name, the category as the problem extraction condition, the keyword, and the like to the operator, but the display unit has a tabular display surface, In the case of the input of a problem for each row of the table, a cell for selecting a condition to be extracted as a problem or a condition not to be extracted, a plurality of keywords including a problem name and a condition, and a plurality of categories defining a plurality of categories When cells are arranged and a category is input, the table has a configuration in which a category name, a plurality of keywords included in the category, and a plurality of cells defining a plurality of categories are arranged.

Next, the input unit of the dedicated rule editor described in the first invention includes a keyword, a category that is a keyword set for collecting keyword fluctuations and synonyms, a problem, a category for diagnosing whether or not it is a problem, and A rule that is a combination of keywords can be input and edited. Therefore, the problems, categories, keywords, rules, etc. displayed on the display unit can be edited, newly added, or changed.

In addition, the input unit has attributes such as a problem name, a category name, a keyword, and a rule in cells other than the edit-inhibited cell display of the table displayed by the display unit in response to the input operation of the free description sentence by the operator. Input editing is possible. This is a feature of the third invention to be patented.

Further, in the dedicated rule editor according to the present invention, a cell for selecting a condition for extracting a problem or a condition for not extracting a problem is arranged in a rule defined in a tabular row unit. This is a feature of the fourth invention to be patented.

Furthermore, in the dedicated rule editor according to the present invention, cells for selecting whether or not to define the keywords and categories in the problem extraction conditions to be defined as rules in tabular rows as the order of appearance are arranged. This is the feature of the fifth invention to be patented.

Furthermore, in the rule definition that does not extract a problem in a tabular row unit in the dedicated rule editor according to the present invention, a keyword and a category defined by the rule for extracting a problem are defined, and a category that is denied to the defined keyword or category A cell for selecting whether or not to add the keyword at the front or back is arranged. This is the feature of the sixth invention to be patented.

Furthermore, the dedicated rule editor according to the present invention is configured to provide a selection operation for reflecting a rule in a tabular row unit or a comment to be recorded as a note. This is a feature of the seventh invention to be patented.

The dedicated rule editor according to the present invention includes a keyword storage unit for storing a defined keyword, a combination of a defined problem and a category that is a problem extraction condition, and a keyword in addition to the input unit and the display unit having the above-described features. , A category and a keyword rule included in the category, a rule storage unit, a keyword storage unit and an inspection unit that verifies whether there is a contradiction between the keywords and rules stored in the rule storage unit, and a keyword stored in the keyword storage unit and the rule storage unit And a conversion unit that converts rules into rules for language analysis.

  Since the dedicated rule editor according to the present invention generates a rule for language analysis only by an input editing operation, its operability is an important factor. However, the dedicated rule editor according to the present invention is a method in which an operator operates as a rule and reflects the data in principle. Therefore, the operator feels that the operability is good when the operator can perform the operation as expected when it is reflected in the data. Accordingly, the present inventors have developed a rule reproduction-dedicated rule editor in consideration of such a viewpoint so that the operator feels easy to operate.

  When extracting a problem from a free description sentence in an electronic medical record, the basic procedure is to extract a keyword that is assumed to be a problem from the free description sentence and to estimate the problem from the combination of the extracted keywords. .

  For this reason, the operator of the dedicated rule editor starts with keywords, keyword notations, keyword synonyms, keywords and keyword notations and keyword synonyms keyword sets, categories, problem names, problem names and problems. A rule that is a combination of a keyword and a category to be estimated is input and edited. For this purpose, it is necessary to always display a keyword, a problem name, a category name, a keyword included in the category, and a rule that is a combination of the keyword and the category for estimating the problem and the problem so that the input and editing can be performed.

  At that time, it is considered that it is easy to understand that a rule is a combination of a keyword and a category for estimating a problem and a problem, and that a keyword or category included in a category is input and edited and displayed as one unit of definition. The description method was changed to the line unit definition method. Since there are usually multiple definitions of rules necessary to extract a problem, there are inevitably multiple lines, and multiple definition item cells are provided in the line, resulting in a tabular display. .

  That is, the display unit according to the present invention is represented by a tabular display surface. In one row of the table, a cell for inputting and displaying a problem name, a keyword and a category for estimating the problem are input and displayed. Cell, a cell for inputting and displaying a category name, and a plurality of cells for inputting and displaying a keyword to be included in the category and a category are configured, and this is a basic style constituting a table.

  When defining a rule on the above-mentioned tabular display surface, there are many cases where a rule that is not estimated as a problem is defined by adding a keyword that is in a rule that is estimated as a problem or a keyword that is denied to a category before and after. In this case, first, a definition that estimates a problem in units of rows in the table is performed, and then a definition that does not estimate a problem by selecting addition of a category to be denied is used to shorten a plurality of operations to one operation, thereby improving operability. The improvement is aimed at.

  In addition, when defining a rule to be estimated as a problem, in the case of a combination of a keyword and a category, the appearance order may be necessary. Therefore, in the case of a combination of a keyword and a category, it is necessary to specify whether the order of appearance is in order or out of order. In such a case, switching between order or random order is provided for each line so that the operation becomes easy.

  Furthermore, there is a case where it is desired to write a memo in order to consider a rule presumed to be a problem. In this case, it is necessary to switch between recording as a memo and reflecting it as a rule (rule). As a result, it is possible to temporarily switch from a memo to a state reflected as a rule (rule), so it is sufficient even if the operator describes it with an idea. This reduces the burden on the operator and improves usability.

  Next, the keyword input in this way is stored in the keyword storage unit, the rule is stored in the rule storage unit, and whether or not there is a contradiction in the rule is verified in the inspection unit. Then, after confirming that there is no contradiction, the conversion unit converts the rule into a language analysis rule and generates a rule. By using a dedicated rule editor, it is possible to easily generate a rule for language analysis from an input operation based on the operator's idea by such a procedure.

  As described above, the dedicated rule editor according to the present invention is a dedicated rule editor that is very easy to operate for the operator, and does not put a lot of load on the operator and is simple operation without repeated use. It is a tool that can generate rules by language analysis of problem extraction.

  The dedicated rule editor according to the present invention generates a rule only by inputting and editing the definition of a rule for extracting a problem by performing linguistic analysis on a free description sentence of a medical finding document. When defining rules for such problem extraction, there is no need for knowledge of program development, no need to think about the logical structure, and no need to learn and learn the definition method, and it is easy and simple to input into the editor. Rules can now be generated.

In particular, in order to perform appropriate problem extraction, add new keywords, add sway keywords, add synonyms, add keywords and category combinations for problem extraction conditions, or add keywords and categories for problem extraction conditions. Even when a detailed response such as correction is required, it can be easily performed by using the dedicated rule editor according to the present invention. As a result, it has become possible to continuously extract problems by analyzing the language from the free description of the findings in the electronic medical record.

  With the present invention, practical use of a method for extracting a problem by analyzing a language from a free description sentence of an electronic medical record progresses, and more advanced and effective use of medical information obtained from an electronic medical record that has not been performed conventionally. You can now.

  As an additional function of electronic medical records and report systems, a program that implements a method for linguistic analysis of observation free description sentences according to rules is incorporated, and when a free description sentence of observation is stored, the problem is extracted and the extracted problem is referenced and linked. In order to support medical practice, the dedicated rule editor according to the present invention is also added, so that the definition of rules can be generated and edited. It is to be able to provide.

Hereinafter, the present invention will be described in detail based on illustrated embodiments.
FIG. 1 is a block diagram showing a configuration of a dedicated rule editor 1 according to the present invention, and FIG. 2 is a block explanatory diagram showing an embodiment when the dedicated rule editor 1 according to the present invention is incorporated in an electronic medical record system. is there.

In the figure, reference numeral 2 denotes an input unit, a keyword, a category that is a keyword set that summarizes keyword notations and synonyms, a problem, a rule that is a combination of a category and a keyword that diagnoses whether or not it is a problem, and It is a means that can input edit operation. In the figure, reference numeral 3 denotes a display unit, which is a means for displaying an input edited problem, rule, category, and keyword. In the figure, reference numeral 4 denotes a keyword storage unit, which is a storage means for storing an input keyword, and is connected to the input unit 2 and the display unit 3. In the figure, reference numeral 5 denotes storage means for storing the input category and rule, and is connected to the input unit 2 and the display unit 3. In the figure, reference numeral 6 denotes means for verifying whether there is any inconsistency in the combination of circulation and conflicting rules, and is connected to the keyword storage unit 4 and the rule storage unit 5. In the figure, reference numeral 7 denotes conversion means for generating stored keywords and rules as rules for language analysis, which are connected to the keyword storage unit 4 and rule storage unit 5.

  As described above, the dedicated rule editor 1 according to the present invention includes the input unit 2, the display unit 3, the keyword storage unit 4, the rule storage unit 5, the inspection unit 6, and the conversion unit 7, and the input unit 2 and the display unit 2. The inspection unit 6 and the conversion unit 7 are all connected to the keyword storage unit 4 and the rule storage unit 5.

  Software that realizes a method for language analysis of free description sentences of medical observation documents according to rules is included as an additional function of electronic medical records and report systems, or as a support system for them, and when a free description sentence of findings is stored, a problem is extracted It is a common usage to support quantitative actions by linking to functions other than electronic medical records and report systems from the extracted problem.

  For example, as shown in FIG. 2, the software of the method for extracting a program from a medical finding document provides a language analysis function 9 as a program or a library to an electronic medical record or report system 8, and the description of the medical observation is expressed in a language by the language analysis function. Analyze and extract the problem. At this time, the problem cannot be extracted unless the rule 10 is defined. The basics of Rule 10 are incorporated in the software of the method of extracting a problem from a medical observation document from the beginning, but the rules are added and modified based on the medical findings described by the doctor who is the user. In order to correct and improve the problem extraction conditions, it is preferable to use the dedicated rule editor 1 according to the present invention. That is, by using the dedicated rule editor 1 according to the present invention, it is possible to easily and continuously update the extraction conditions of a problem by performing linguistic analysis from a free description sentence of the findings of an electronic medical record. The practicality of the software for extracting programs by document has improved significantly.

  First, an example of newly generating a rule for extracting “Problem 120” from a free description sentence in a medical finding document will be described. FIG. 3 is an image explanatory diagram of an example in which “left ventricular chamber enlargement” is input as a free description in the medical finding document, the language analysis program analyzes according to the rules, and “problem 120” is extracted. FIG. 4 is a diagram schematically showing the definition portion on the tabular display surface of the dedicated rule editor 1. That is, as shown in FIG. 4, the rule is that the problem 120 "is extracted with the keyword" left ventricular chamber expansion ".

Next, a case where a term for guessing a problem is input to the medical finding document, but no problem is extracted because there is no rule, and a case where a new rule is additionally generated are shown.
FIG. 5 is an image explanatory diagram showing an example in which “LV expansion” is input to the medical finding document and the language analysis program analyzes according to the rule but does not extract the problem because there is no rule. FIG. 6 is a diagram schematically showing an example in which the category “LV” is created and the keyword “LV” is defined in the definition portion on the tabular display surface of the dedicated rule editor 1. is there.
FIG. 7 is a diagram schematically showing an example in which a category with the category name “enlarged” is created and the keyword “enlarged” is defined in the definition portion on the tabular display surface of the dedicated rule editor.
Further, FIG. 8 is a definition part on the table-type display surface of the dedicated rule editor. An example in which a combination rule for specifying a problem 120 on a new line and defining a category LV and a category expansion and extracting the problem 120 is added. FIG.

  Therefore, in order to add “LV expansion” to the extraction condition of the problem 120 (left heart failure), the dedicated rule editor 1 is activated, and “LV” is used as the category name and “LV” as the keyword included in the category as shown in FIG. As shown in FIG. 7, enter “enlarge” as the category name, enter “enlarge” as the keyword included in the category, and enter the categories “LV” and “enlarge” into the problem 120 (left heart failure) as shown in FIG. To do. This saves the added definition and completes the rule generation.

Next, a case will be described in which the rule of the language analysis program for the medical finding document is updated to the generated rule. FIG. 9 is an image view of an example in which after the rules are updated by the dedicated rule editor 1, the “LV expansion” described in the findings of the electronic medical record is analyzed by the language analysis program and “Problem 120” is extracted. It is. This is an example of extracting the problem 120 by analyzing the free description “LV expansion” after the update.

  Furthermore, the case where the problem of a synonym cannot be extracted from the term described in the free description sentence of a medical finding document is shown. For example, the problem 120 (left heart failure) may not be extracted as a result of describing and analyzing “left ventricular expansion” in the free description sentence of the medical finding document. Therefore, a case where the synonym “left ventricular expansion” is added to the extraction condition of the problem 120 (left heart failure) is shown.

FIG. 10 is an image explanatory diagram of an example in the case where the “left ventricular expansion” described in the medical finding document is analyzed by the language analysis program and the problem cannot be extracted.
FIG. 11 is a diagram schematically showing an example in which “left chamber” is added to the 02 column of the category name “LV” row in the definition portion on the tabular display surface of the dedicated rule editor 1.

  Therefore, in order to add “left ventricular expansion” to the extraction condition of the problem 120 (left heart failure), the dedicated rule editor 1 is started, and the keyword “LV” in the category “LV” as shown in FIG. Add “room”. Save the added definition and create a rule.

Next, a case where the rule of the language analysis software of the medical finding document is updated to a newly generated rule will be described.
FIG. 12 is an image-like example of the case where after the rule is updated by the dedicated rule editor 1, “left ventricular expansion” described in the findings of the electronic medical record is analyzed by the language analysis program and “problem 120” is extracted. FIG. This is an example of extracting the problem 120 by analyzing the free description sentence “left ventricular expansion” after the update.

  When the dedicated rule editor 1 according to the present invention is used, it becomes possible to add a keyword such as a notation swing or a synonym spirally on the spot, and it is defined in the past by the operator. It is possible to learn naturally by referring to keyword notation fluctuations and synonyms, and it is possible to maintain the sustained effectiveness of the language analysis function by the flexibility of the rules.

As described above, the dedicated rule editor 1 according to the present invention was developed as a tool that can generate linguistic analysis rules by a simple free description input editing operation. From the electronic medical chart system and the medical observation report system, It is a medical exercise support tool that aims to make high-level use of the medical information obtained. For this reason, it is suitable as a rule generation support tool used in the “problem extraction engine” incorporated in medical risk warning systems and clinician decision support systems, etc., and it incorporates a language analysis function for free description sentences of medical observation documents It can also be used as a maintenance tool for analysis rules.

It is a block diagram which shows the structure of the exclusive rule editor which is this invention. It is a block diagram which shows the Example at the time of incorporating the exclusive rule editor which is this invention in an electronic medical chart system. It is an image explanatory diagram of an example when a free description sentence “left ventricular chamber expansion” is input to a medical finding document, a language analysis program analyzes according to a rule, and “problem 120” is extracted. It is the figure which showed typically the definition part in the display surface of the table format of a special rule editor. FIG. 5 is an image explanatory diagram illustrating an example in which “LV expansion” is input to the medical finding document and the language analysis program performs analysis according to the rule but does not extract the problem because there is no rule. FIG. 6 is a diagram schematically showing an example in which the category “LV” is created and the keyword “LV” is defined in the definition portion on the tabular display surface of the dedicated rule editor. FIG. 7 is a diagram schematically showing an example in which a category with the category name “enlarged” is created and the keyword “enlarged” is defined in the definition portion on the tabular display surface of the dedicated rule editor. FIG. 8 is a definition part on the table-type display surface of the dedicated rule editor. An example in which a combination rule for specifying a problem 120 on a new line and defining a category LV and a category expansion and extracting the problem 120 is added. It is the figure shown typically. FIG. 9 is a conceptual explanatory diagram of an example in which “LV expansion” described in the findings of the electronic medical record is analyzed by a language analysis program after “Rule 120” is extracted after updating the rule by the dedicated rule editor. . FIG. 10 is an image explanatory diagram of an example in the case where the “left ventricular expansion” described in the medical finding document is analyzed by the language analysis program and the problem cannot be extracted. FIG. 11 is a diagram schematically showing an example in which “left room” is added to the 02 column of the category name “LV” row in the definition portion on the tabular display surface of the dedicated rule editor. FIG. 12 is a conceptual explanatory diagram of an example in which “left ventricular expansion” described in the findings of the electronic medical record is analyzed by a language analysis program after extracting the “problem 120” after updating the rules with the dedicated rule editor. is there.

Explanation of symbols

1: exclusive rule editor 2: input unit 3: display unit 4: inspection unit 5: conversion unit 6: keyword storage unit 7: rule storage unit 8: electronic medical record and report system 9: language analysis function 10: rule (dedicated rule editor output)

Claims (7)

In the method of extracting medical problems such as disease name, symptom, examination, treatment, treatment, etc., according to the rules for language analysis of free description sentences in medical observation documents, rules are generated by inputting and editing the definition of rules for language analysis. A dedicated rule editor,
An input unit that inputs and edits a keyword, a category that is a keyword set that summarizes keyword notations and synonyms, a problem, a category that diagnoses whether or not it is a problem, and a rule that is a combination of keywords;
A display section for displaying the input and edited problems, rules, categories, and keywords;
A keyword storage unit connected to the input unit or the display unit and storing the input keyword;
A rule storage unit that is connected to the input unit or the display unit and stores an input category and rule,
An inspection unit that is connected to the keyword storage unit and the rule storage unit, and verifies whether there is a contradiction in the combination of circulation and conflicting rules,
A dedicated rule editor comprising a conversion unit that is connected to the keyword storage unit and the rule storage unit and generates the stored keywords and rules as rules for language analysis. The display section has a tabular display surface, and for each row of the table, a cell for inputting and displaying a problem name, a cell for inputting and displaying a keyword and a category for estimating the problem, and a category name are input and displayed. The basic style is to provide multiple cells for entering and displaying keywords and categories to be included in cells and categories.
When inputting the problem, a cell for selecting a condition to be extracted as a problem or a condition not to be extracted, and a plurality of cells defining a problem name, a plurality of keywords as a condition, and a plurality of categories are arranged,
In the case of inputting a category, a plurality of cells defining a category name, a plurality of keywords included in the category, and a plurality of categories are arranged.
The dedicated rule editor according to claim 1, wherein the dedicated rule editor is configured in a table format. Corresponding to the input operation of the free description sentence of the operator, the input part which can input and edit attributes such as the problem name, category name, keyword, rule, etc. in cells other than the edit prohibited cell of the table displayed on the display part. The dedicated rule editor according to claim 1, further comprising: The dedicated rule editor according to claim 1, wherein a cell for selecting whether a condition for extracting a problem or a condition for not extracting is arranged in a rule defined in units of rows in a table format. 2. The dedicated rule editor according to claim 1, wherein cells for selecting whether or not to define the keywords and categories as the problem extraction conditions to be defined as rules in a tabular row are in the order of appearance. . In the definition of rules that do not extract problems in tabular rows, define the keywords and categories that are defined in the rules that extract problems, and add the keywords in the negative category to the keywords or categories that you define before or after 2. The dedicated rule editor according to claim 1, wherein a selection cell of whether or not to perform is arranged. 2. The dedicated rule editor according to claim 1, wherein a selection operation is provided for selecting whether to reflect in a tabular line unit as a rule or to make a comment that a record is written as a memo.