JP2013109475A - Document analysis system, document analysis method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for accurately estimating preferential correction points and quality of a document including ambiguous words in consideration of the influence of the ambiguous word on the document quality in each use scene.
A document analysis system according to the present invention searches an ambiguous word in a document based on word information of each word used in a sentence constituting an externally input document, and there is an ambiguous word. Is an ambiguous word search unit that extracts the ambiguous word, an estimated key word extraction unit that extracts each estimated key word for each ambiguous word from the document based on an estimated key word extraction rule for extracting the estimated key word; A concept scale estimation unit that estimates a concept scale index, which is the scale of the concept of the estimated key word extracted by the estimation key word extraction unit based on the concept scale estimation rule, and each ambiguous word based on the concept scale index. An ambiguity calculation unit that calculates ambiguity information for each combination with the estimated key word.
[Selection] Figure 1

Description

  The present invention relates to a document analysis system, a document analysis method, and a program, and more particularly to a document analysis system, a document analysis method, and a program for estimating preferential correction points and quality of a document.

  2. Description of the Related Art In recent years, a technology has been developed that uses an information processing apparatus to analyze a document written in a natural language and analyze preferential correction points and quality of the document. An example of the document analysis technique is disclosed in Patent Document 1 as “Japanese sentence correcting device, Japanese sentence correcting method, and Japanese sentence correcting program”.

  The Japanese sentence correction device disclosed in Patent Document 1 includes a sentence composition database in which rules relating to structures that are easy to machine translate are registered, and a script / term database in which rules relating to letters and terms that are easy to machine translate are registered. And a style database in which rules relating to styles that are easy to machine translate are registered, a correction location extraction means, a display means, a correction means, and an output means. The similar expression extraction device having such a configuration operates as follows.

  The correction part extraction means relates to the above-mentioned rules such as “create a sentence by the number of predicates” from the original Japanese text to be read, and the above-mentioned script / terms such as “the sentence does not include any characters, typographical errors, or omissions”. A configuration, script, term, or style that violates a rule or a rule relating to the style such as “the sentence does not include an ambiguous phrase” is extracted.

  The display means displays the extracted configuration, script / term or style.

  The correcting means corrects the original Japanese sentence by correcting, adding, or deleting the extracted configuration, script, term, or style using data input from the outside.

  The output means outputs an easy-to-translate sentence obtained by the correction.

  With this configuration, points that violate the above rules are extracted and displayed from the original Japanese text, and corrections such as external correction, addition, and deletion are supported.

  Further, other examples of the document analysis technique are disclosed in Patent Documents 2 to 5.

  In Patent Document 2, the concept of information managed in a database is registered in an ontology in an array according to the abstraction level of each concept, and a variety of databases are frequently added or deleted. Even if it exists, the technique which becomes easy to maintain and manage the database selection menu is disclosed.

  Patent Document 3 discloses a technique that enables an appropriate search by considering the concept hierarchy between words in the search.

  Patent Document 4 discloses a technique that allows a user to create a sentence more easily by predicting and displaying a word to be input by referring to a database from words positioned before the first reading information of the word to be input next. Disclose.

  Patent Document 5 discloses a machine for automatic translation, automatic abstract creation processing, etc. by embedding information (additional information) useful for mechanical processing in an arbitrary character string in a text document being created. Disclosed is a technique capable of creating a text document suitable for an intelligent process.

JP 2007-316834 A JP 2000-222436 A JP 2003-022277 A JP 2005-234742 A JP-A-11-282844

  The problem of the techniques described in Patent Documents 1 to 5 is that the accuracy of preferential correction and quality estimation of documents including ambiguous words such as “etc.”, “etc.”, “toka” or “other than” is low. is there.

  The reason for this is that ambiguous words that have the effect of causing the reader to infer the abbreviated concepts such as “etc.”, “etc.”, “toka” or “other than” are used in the text to improve the quality of the document. This is because the magnitude of the effect is different. Therefore, the method for detecting the presence or absence of a registered ambiguous word as in the technique described in Patent Document 1 cannot take into account the magnitude of the influence of the ambiguous word on the quality of the document in each use scene.

  Note that the quality of the document here is a feature that includes at least unambiguous characteristics such as difficulty of understanding between the writer and reader of the document and ease of understanding for the reader of the document. This refers to the characteristics related to the performance of document information transmission.

  As described above, the object of the present invention is to accurately estimate preferential correction points and quality of a document including ambiguous words in consideration of the magnitude of the influence of ambiguous words on the quality of the document in each use scene. An object of the present invention is to provide a document analysis system, a document analysis method, and a program that can be used.

In order to achieve the above object, the document analysis system according to the present invention is based on word information of each word used in a sentence constituting a document input from the outside, and omits the omitted concept in the document. An ambiguous word, which is a word that has an action to be estimated by the reader from the context, is searched. Estimated key word that extracts each estimated key word for each ambiguous word from a document based on an estimated key word extraction rule for extracting an estimated key word that is a word that becomes a clue when a reader estimates the concept A concept for estimating a concept scale index, which is a scale of the concept of the estimated key word extracted by the estimated key word extraction section, based on a concept scale estimation rule for estimating the scale of the concept of the estimated key word Rule An ambiguity calculation unit that calculates ambiguity information for each combination of each of the ambiguous words and the estimated key word based on the concept scale index. The document analysis method is a word having an action of causing a reader to estimate an omitted concept from a context based on word information of each word used in a sentence constituting a document inputted from the outside. A certain ambiguous word is searched, and when there is an ambiguous word, the ambiguous word is extracted, and an estimated key word that is a clue when a reader estimates a concept omitted by each extracted ambiguous word is extracted. Based on the estimated key word extraction rule for extracting each estimated key word for each after ambiguity from the document and extracting based on the concept size estimation rule for estimating the size of the concept of the estimated key word Estimated Estimates the concept scale index is a scale word concept, on the basis of the concept scale index is calculated ambiguity information the every combination of the ambiguous word and the estimated key word.

  In order to achieve the above object, the program according to the present invention is based on the word information of each word used in the sentence constituting the document input from the outside, and omits the omitted concept in the document from the context. Search for ambiguous words, which are words that have an action to be estimated by the reader. If there is an ambiguous word, the ambiguous word is extracted, and this is a clue when the reader estimates a concept omitted by each extracted ambiguous word. Based on an estimated key word extraction rule for extracting an estimated key word that is a word, each estimated key word for each after ambiguity is extracted from the document, and a concept scale for estimating the concept size of the estimated key word Based on the estimation rule, a concept scale index that is the scale of the concept of the extracted estimated key word is estimated, and based on the concept scale index, ambiguity information is obtained for each combination of each ambiguous word and the estimated key word. And out, to perform the processing to computer.

  According to the present invention, it is possible to accurately estimate preferential correction points and quality of a document including ambiguous words in consideration of the magnitude of influence of the ambiguous word on the document quality in each use scene. it can.

It is a block diagram which shows the structure of the document analysis system 100 which concerns on Embodiment 1 of this invention. 2 is a diagram illustrating a hardware configuration of a document analysis system 100. FIG. It is a flowchart which shows the whole operation | movement of the document analysis system 100 which concerns on Embodiment 1 of this invention. It is a block diagram which shows the structure of the document analysis system 200 which concerns on Embodiment 2 of this invention. It is a block diagram which shows the structure of the document analysis system 100 in an Example. It is explanatory drawing which shows the example of a part of ambiguous word Waj extracted from an example sentence. It is explanatory drawing which shows the example of a part of presumed key word Wbj when the word which existed just before the ambiguous word Waj is made into presumed key word Wbj. It is explanatory drawing which shows the example of a part of presumed key word Wbj when the word which has a dependency relation with the word which exists immediately before the ambiguous word Waj is made into presumed key word Wbj. It is explanatory drawing which shows the example of the conceptual scale parameter | index Ibj of the presumed key word Wbj of FIG. It is explanatory drawing which shows the example of the conceptual scale parameter | index Ibj of the presumed key word Wbj of FIG. It is explanatory drawing which shows the example of the ambiguity information Aj based on the conceptual scale parameter | index Ibj of FIG. It is explanatory drawing which shows the example of the ambiguity information Aj based on the conceptual scale parameter | index Ibj of FIG.

[Embodiment 1]
First, Embodiment 1 of the present invention will be described with reference to FIGS.

  FIG. 1 is a block diagram showing a configuration of a document analysis system 100 according to Embodiment 1 of the present invention.

  Referring to FIG. 1, a document analysis system 100 according to Embodiment 1 of the present invention includes a document input unit 10, a word analysis unit 20, an ambiguous word search unit 30, an estimated key word extraction unit 40, a concept scale estimation unit 50, an ambiguous word. A sex calculation unit 60, an ambiguity output unit 70, an ambiguous word database 110, and a concept database 120 are included. The document analysis system 100 may include each of the above-described components in a single electronic device or a system including a server and an electronic device and an information communication network such as the Internet that connects these components to each other.

  The document analysis system 100 is a document analysis system for estimating preferential correction points and quality of a document in consideration of the magnitude of the influence of the ambiguous word on the quality of the document in each use scene for the document including the ambiguous word. is there. Here, the “ambiguous word” refers to a word having an effect of causing a reader to estimate an abbreviated concept such as “etc.”, “etc.”, “toka” or “other than” from the context.

  When the document analysis system 100 is configured with a single electronic device, the document analysis system 100 can be realized by a computer that operates under program control.

  FIG. 2 is a diagram showing a hardware configuration when the document analysis system 100 is realized by a single electronic device. As shown in FIG. 2, the document analysis system 100 includes an input device 1 that inputs data, a data processing device 2, an output device 3 that outputs processing results in the data processing device 2, and a storage device 4 that functions as various databases. Including.

  The data processing device 2 stores in the RAM 6 in accordance with a read only memory 5 (ROM 5) for storing a program, a random access memory 6 (RAM 6) used as a work area for temporarily storing data, and a program stored in the ROM 5. A central processing unit 7 (CPU 7) for processing the processed data.

  The input device 1 constitutes a part of the document input unit 10. The data processing device 2 operates as the document input unit 10, the word analysis unit 20, the ambiguous word search unit 30, the estimated key word extraction unit 40, the concept scale estimation unit 50, and the ambiguity calculation unit 60. It operates as the ambiguous word database 110 and the concept database 120, and the output device 3 may operate as the ambiguous output unit 70.

  Next, each component constituting the document analysis system 100 will be described in more detail.

  The document input unit 10 receives document data. The document that the document input unit 10 accepts as data may be any document, but is basically a document that may include ambiguous words and is subject to preferential correction points and quality analysis. . Further, the document referred to here may be a document group that is a bundle of a plurality of documents. Specifically, the document input unit 10 includes an input device 1 and a CPU 7. When the input device 1 inputs a document, the CPU 7 stores the document data in the RAM 6.

  The word analysis unit 20 extracts word information, which is information about words used in sentences constituting the document, from the document (document data) input by the document input unit 10. This extraction is executed based on program control by the CPU 7 of the data processing unit. The word analysis unit 20 may extract word information of all words used in each sentence by applying, for example, morphological analysis or syntax analysis to the sentences constituting the input document.

  Here, the word may include not only a self-supporting independent word such as a noun, a verb, and an adjective, but also a particle. The word information includes information on the word itself being used. The word information may include information regarding the location of the word. The information related to the position of the word may be information that can identify the use location, and corresponds to the order of appearance of the sentence in which the word exists, the page, the chapter, the section, and the item on the table of contents.

  Further, the word information may include a part of speech, a case, a particle to be combined, a dependency relationship between words, and the like as necessary.

  The ambiguous word database 110 stores ambiguous words. The ambiguous word database 110 may be a database that searches and responds to an inquiry from an ambiguous word search unit 30 to be described later regarding a specific word as to whether or not the word to be inquired is registered as an ambiguous word. The ambiguous word database 110 may be a database on the Internet.

  The ambiguous word search unit 30 refers to the ambiguous word database 110 and searches for ambiguous words based on the word information extracted by the word analysis unit 20, that is, word information of all words used in each sentence. . Furthermore, when the ambiguous word is found in the ambiguous word database 110 as a result of the search, the ambiguous word search unit 30 extracts the ambiguous word. In addition, the ambiguous word search unit 30 extracts information on the position of the ambiguous word, which is information on the appearance order and page of the sentence including the ambiguous word, the chapter, the section, and the item on the table of contents together with the ambiguous word. May be.

The estimated key word extraction unit 40 extracts an estimated key word from the document for each ambiguous word extracted by the ambiguous word search unit 30 based on the estimated key word extraction rule. An estimated key word is a word that is used as a clue when a reader estimates a concept omitted by an ambiguous word. Here, an estimated key word extraction rule is used when a reader estimates a concept omitted by an ambiguous word. Any rule may be used as long as it is a rule for extracting a clue word. The estimated key word extraction rule may be, for example, a rule in which a word existing immediately before an ambiguous word or a word modified by an ambiguous word (modified word) is an estimated key word. The presumed key word extraction rule may be a rule in which a word corresponding to any of a subject or a predicate of a sentence including an ambiguous word or a word having a dependency relationship with a word existing immediately before the ambiguous word is used as the presumed key word. . Alternatively, the estimated key word extraction rule may be a rule in which a combination of words that are the estimated key words described above is an estimated key word.

  The concept database 120 stores the concept classification of the collected words and general concept information such as general synonyms, synonyms, and usage. The concept database 120 may be a database that searches for general concept information of a word and responds to an inquiry about a specific word. The concept database 120 may be a thesaurus or the like in which words are classified and organized according to the upper and lower relationships of words, partial relationships and overall relationships, synonymous relationships, or similar relationships. The concept database 120 may be a database on the Internet.

  The concept scale estimation unit 50 receives the estimation key word extracted by the estimation key word extraction unit 40 and refers to the concept database 120, and general concept information related to the estimation key word, and a predetermined concept size estimation. Based on the rule, a concept scale index that is the scale of the concept of the estimated key word is estimated.

  Here, the “concept scale estimation rule” for estimating the concept scale index may be a rule that can estimate the relative size of the concept of the estimated key word. For example, the concept scale estimation rule may be a rule that uses the number of synonyms or the number of synonyms as a concept scale index.

  Alternatively, when a concept vector is created using a thesaurus in which concepts at a plurality of levels are registered as general concept information in the concept database 120, the concept scale estimation rule uses a value that is in a monotonically increasing relationship with the level of the hierarchy. A rule may be used as a scale index. Specifically, the concept scale estimation rule uses the hierarchy on the highest concept classification to which each estimation key word belongs, and is 5 for the highest classification, 4 for the middle classification, Set to 3 for words, 2 for words, and 1 for no registration. As described above, the concept scale estimation rule may be a rule that uses a value that is in a monotonically increasing relationship with the height of the hierarchy as a concept scale index. Alternatively, the concept scale estimation rule may be a rule using the total number of words registered in the subclass of the concept class to which each estimated key word belongs as a concept scale index.

  The ambiguity calculation unit 60 calculates ambiguity information for each combination of the estimated key word and the ambiguous word based on the concept scale index estimated by the concept scale estimation unit 50.

  Here, the ambiguity information may be an ambiguity index that is an index related to ambiguity for each combination of each estimated key word and ambiguous word. For example, the ambiguity index is an index related to ambiguity that is derived by a function that has a monotonically increasing relationship with the concept scale index. In addition to the ambiguity index, the ambiguity information may include information on the estimated key word and the position of the ambiguous word in the input document. Alternatively, the ambiguity information may be an index that represents the ambiguity index for each combination of the estimated key word and the ambiguous word in the entire document or in an arbitrary range and represents the quality of the document.

  The ambiguity output unit 70 outputs the ambiguity information calculated by the ambiguity calculation unit 60. The output form may be output in a required form. For example, the ambiguity output unit 70 performs processing such as color coding, emphasis by bold or enlargement of characters on the combination of each estimated key word and ambiguous word in the document, and the entire document subjected to the processing is processed. It may be output.

  Alternatively, the ambiguity output unit 70 may output a table in which combinations of estimated key words and ambiguous words are extracted. The ambiguity output unit 70 may output only the combination of each estimated key word and ambiguity word whose ambiguity index is larger than the arbitrarily set threshold value. The ambiguity output unit 70 may perform color coding based on the ambiguity index, emphasis by bold characters, or imparting strength to the character size of words.

  When an ambiguity index for each combination of each estimated key word and ambiguous word is an index obtained by tabulating the entire document or an arbitrary range, and an index indicating the quality of the document is output as ambiguity information, the ambiguity output unit 70 The ambiguity information may be output in a table format.

  Alternatively, the ambiguity output unit 70 may be configured to accept selection of each output form by the user and shift from the base display form to a table as necessary. Further, the ambiguity output unit 70 may selectively output a verb, a noun, or the like as necessary.

  The ambiguity output unit 70 may output ambiguity information and accept correction of text by the user.

  Next, the overall operation of the document analysis system 100 according to the first embodiment of the present invention will be described in detail with reference to the flowchart of FIG. Note that the flowchart shown in FIG. 3 and the following description are examples of processing, and the processing order and the like may be changed or the processing may be returned or repeated depending on the processing that is appropriately obtained.

  The document input unit 10 receives an input of a document to be analyzed (step A1).

  The word analysis unit 20 extracts word information of all words used in each sentence by applying morphological analysis and syntax analysis to each sentence constituting the document (step A2).

  The ambiguous word search unit 30 refers to the ambiguous word database 110 for the presence / absence of ambiguous words in the document based on the word information of all the words used in each sentence extracted by the word analyzing unit 20, and the ambiguous words are found. If there is, the ambiguous word is extracted (step A3).

  The estimated key word extraction unit 40 extracts, as estimated key words, words that are clues when the reader estimates a concept omitted by the ambiguous words based on the estimated key word extraction rules for each ambiguous word (step A4).

  The conceptual scale estimation unit 50 refers to the conceptual database 120 for the general conceptual information of the estimated key word, and based on the obtained general conceptual information and a predetermined conceptual scale estimation rule, the conceptual scale that is the conceptual scale of the estimated key word An index is estimated (step A5).

  The ambiguity calculation unit 60 calculates ambiguity information for each combination of each estimated key word and ambiguous word based on the concept scale index estimated by the concept scale estimation unit 50 (step A6).

  The ambiguity output unit 70 outputs the ambiguity information calculated by the ambiguity calculation unit 60 (step A7).

As described above, according to the document analysis system 100 according to the first embodiment, with respect to a document including an ambiguous word, the magnitude of the influence of the ambiguous word on the quality of the document in each use scene is considered, and the document is accurately obtained. Preferential correction points and quality can be estimated. ,
The reason is that the estimated key word extraction unit 40 extracts a word that becomes a clue when the reader estimates a concept omitted by an ambiguous word in the document as an estimated key word, and the concept size estimation unit 50 uses the concept of the estimated key word. This is because the concept scale index, which is the scale of, is estimated. By calculating the ambiguity index for each combination of the estimated key word and the ambiguous word, the document analysis system 100 can perform analysis in consideration of how the ambiguous word is used in the sentence that affects the quality of the document. Become.

[Embodiment 2]
Next, Embodiment 2 of the present invention will be described with reference to FIG.

  FIG. 4 is a block diagram showing a configuration of the document analysis system 200 according to Embodiment 2 of the present invention.

  Referring to FIG. 4, the document analysis system 200 according to the second exemplary embodiment of the present invention includes an ambiguous word search unit 35, an estimated key word extraction unit 45, a concept scale estimation unit 55, and an ambiguity calculation unit 65.

  The ambiguous word search unit 35 has a function of causing a reader to estimate an omitted concept from a context based on word information of each word used in a sentence constituting the document input from the outside. If there is an ambiguous word, the ambiguous word is extracted.

  The estimated key word extraction unit 45 extracts an estimated key word extraction rule for extracting an estimated key word that is a word that becomes a clue when the reader estimates a concept omitted by each ambiguous word extracted by the ambiguous word extraction unit 35. Based on the above, each estimated key word for each ambiguous word is extracted from the document.

  The conceptual scale estimation unit 55 calculates a conceptual scale index that is the conceptual scale of the estimated key word extracted by the estimated key word extraction unit 45 based on a conceptual scale estimation rule for estimating the conceptual scale of the estimated key word. presume.

  The ambiguity calculation unit 65 calculates ambiguity information for each combination of the estimated key word and the ambiguous word based on the concept scale index estimated by the concept scale estimation unit 55.

  As described above, according to the document analysis system 200, a document including an ambiguous word is preferentially corrected with high accuracy in consideration of the magnitude of the influence of the ambiguous word on the document quality in each use scene. Points and quality can be estimated.

  The document analysis method according to the present invention is an operation method executed by the document analysis system 100 or the document analysis system 200 described above. The program according to the present invention may be a program that causes a computer to execute the operations described in the first or second embodiment.

[Example]
Next, the operation of the document analysis system 100 according to the first embodiment of the present invention will be described using specific examples with reference to FIGS.

  The purpose of this embodiment is as follows.

  First, the document analysis system 100 sets an ambiguity index IA that is an ambiguity that the ambiguous word Wa gives to the quality of the document D in the document D that should exclude ambiguous parts, such as proposals and specifications related to information system construction. Calculation is performed for each use scene of each ambiguous word Wa. The ambiguous word Wa is a word having an action of causing a reader to estimate an abbreviated concept such as “etc.”, “etc.”, “toka” or “other than” from the context.

  Based on the calculated ambiguity index IA, the document analysis system 100 estimates ambiguity information A related to the position of each ambiguous word in the document and the degree of ambiguity. The document analysis system 100 outputs the estimated ambiguity information A, so that the ambiguity portion to be corrected with priority in the document D can be grasped, or a low-quality document can be clarified as a comparison between a plurality of documents. Facilitates and streamlines document improvement.

  In this embodiment, the document analysis system 100 includes a document analysis system 300 and an Internet server 400 as shown in FIG.

  The document analysis system 300 operates on the PC terminal 500 possessed by the analyst B, and through the input unit 600 and the output unit 700, the analysis practitioner B wants to estimate the ambiguity index A. Realize input and presentation of ambiguity information A.

  The Internet server 400 is connected to the PC terminal 500 of the analysis person B who has implemented the document analysis system 300 via a communication network. The Internet server 400 enables a search for presence / absence of registration as an ambiguous word Wa in response to a word inquiry from the document analysis system 300. Further, the Internet server 400 enables searching of general concept information Cg related to word concept classification, general ambiguous words, synonyms, and usage in response to a query of concept information such as word meaning.

  Next, the correspondence between FIG. 5 and FIG. 1 will be described.

  The document input unit 10, the word analysis unit 20, the ambiguous word search unit 30, the estimated key word extraction unit 40, the concept scale estimation unit 50, and the ambiguity calculation unit 60 are included in the document analysis system 300. The ambiguity output unit 70 operates as the output unit 700. The ambiguous word database 110 and the concept database 120 are included in the Internet server 400.

  The document analysis system 300 and the Internet server 400 including such components operate as follows.

  The document analysis system 300 accepts an input of a document D that the analysis operator B wants to obtain ambiguity information A for eliminating an ambiguous part, such as a proposal or specification regarding information system construction, via the input unit 600. . The document analysis system 300 applies morphological analysis and syntax analysis to each sentence constituting the document D, decomposes it into words constituting the document, analyzes the dependency relationship and case, and thereby analyzes the word W and its case. A dependency structure or the like is extracted as word information. Here, the “word” may include an expression including a plurality of words such as “related to”. Furthermore, the document analysis system 300 assigns sentence numbers as serial numbers in the order of appearance of the sentences in the document, and adds the sentence numbers of sentences including each word to the word information.

  The Internet server 400 stores ambiguous word information Ca, which is information about ambiguous words and ambiguous expressions that have the effect of causing the reader to infer the omitted concepts from the context. In addition, the Internet server 400 also has a function such as a search engine that extracts information on arbitrary words and expressions, so that the query target word exists in the ambiguous word information Ca in response to a query from the document analysis system 300. It may be determined whether or not to present.

  The ambiguous word to be extracted may be any ambiguous word as long as it has the effect of allowing the reader to infer the omitted concept from the context. The ambiguous word corresponds to a word such as “etc.”, “etc.”, “toka”, “other than”, “like”, “like”, and the like. The ambiguous word may also include an expression made up of a plurality of words combined with particles such as “like”, “related”, “related to”, “related to” and the like.

  The document analysis system 300 determines whether or not all words Wi (i = 1, 2,..., N) included in the document D are words corresponding to the ambiguous word of the ambiguous word information Ca in the Internet server 400. Inquire. The document analysis system 300 sets the word Wj (j = 1, 2,..., M) determined to be an ambiguous word as the ambiguous word Waj (j = 1, 2,..., M). And the sentence number of the sentence in which the ambiguous word exists. Note that even when the same ambiguous word is used in the document D multiple times, the usage scenes are different, and the document analysis system 300 extracts the same ambiguous word separately.

  For example, a part of document D is “This purchasing system is aimed at improving the efficiency of logistics operations between domestic bases and overseas branches. The developer should satisfy the requirements of the related documents etc. for the performance of this system. If there are any changes in the requirements, we will discuss each case.If there are any changes, we will describe the changes in the minutes, etc., and develop the developed product according to this description. As a new process, it will be necessary to convert the coefficient of the book value calculation formula from the old version value in order to comply with the new law as a new process. There is no major change in the functional aspects other than these processes. " As shown in FIG. 6, the document analysis system 300 assigns a sentence number to each sentence, and for a sentence in which an ambiguous word exists, the ambiguous word Waj is described corresponding to the sentence number of each sentence. An ambiguous word may be extracted by creating a table.

  The document analysis system 300 extracts a word existing immediately before the ambiguous word Waj as an estimated key word Wbj that becomes a clue when the reader estimates a concept omitted by the ambiguous word. For example, in the case of FIG. 6, the document analysis system 300 extracts words as shown in the rightmost column of FIG. 7 as the estimated key word Wbj.

  The reason why it is effective to use the word existing immediately before the ambiguous word Waj as the estimated key word Wbj is that when there is a notation “document etc.”, the reader estimates other things similar to “document”. It is to become.

  Alternatively, the document analysis system 300 may extract a word having a dependency relationship with the word existing immediately before the ambiguous word Waj as the estimated key word Wbj. For example, in the case of FIG. 6, the document analysis system 300 extracts words as shown in the rightmost column of FIG. 8 as the estimated key word Wbj.

  The reason why it is effective to use the word having a dependency relationship with the word existing immediately before the ambiguous word Waj as the estimated key word Wbj is that there is a notation of “convert coefficients from old version values”, This is because the reader estimates that the operation of “converting” is appropriate.

  Note that at least one estimated key word Wbj may be extracted for the ambiguous word Waj, but the document analysis system 300 may extract a plurality of estimated key words Wbj for one ambiguous word.

  The Internet server 400 stores the general concept information Cg of the thesaurus by classifying and organizing words according to general upper and lower relations, partial relations and whole relations, synonymous relations or synonymous relations of words. Further, the Internet server 400 also has a function of a search engine or the like that extracts information of an arbitrary word, so that general concept information Cg may be extracted and presented in response to an inquiry from the document analysis system 300. . Specifically, the Internet server 400 uses the general concept classification of the query target word as a classification such as “major classification”, “medium classification”, “small classification”, and “phrase” and the words belonging to the classification. The number may be extracted and presented as general concept information Cg.

  The document analysis system 300 inquires the Internet server 400 about the general concept information Cg of the estimated key word Wbj. The document analysis system 300 extracts the highest concept classification to which each estimated key word Wbj belongs from the classification system of the thesaurus general concept information Cg stored in the Internet server 400. The document analysis system 300 extracts the number of words registered in the subclass of the concept class as the value of the concept scale index Ibj.

  For example, when the presumed key word Wbj “document” is classified as “small category” in the classification system of the general concept information Cg, the word “phrase” that is a subordinate category and the word registered as its synonym are extracted. To do. For example, “letter”, “document”, “document”, “official document”, “private document”, “private document”, “contract” are terms and synonyms that are subclasses of the presumed key word Wbj “document”. It is assumed that there are a total of 156 words such as “call”, “memorandum”,. In this embodiment, the document analysis system 300 sets the value of 156 as the value of the conceptual scale index Ibj.

  Further, when the estimated key word Wbj is classified as “word / phrase” in the classification system of the general concept information Cg as in “minutes”, the document analysis system 300 sets the concept scale index because there is no word belonging to the lower level. It may be 1. Alternatively, the document analysis system 300 may extract words registered as synonyms and synonyms of “minutes” from the general concept information Cg, and the number of extracted words may be used as a concept scale index. For example, if three words “memo”, “record”, and “meeting minutes” are obtained as synonyms or synonyms of “minutes”, the document analysis system 300 may set the value of the concept scale index Ibj to 3. . In the case of a word that does not exist in the classification system of the general concept information Cg, there is no word to which it belongs, so the document analysis system 300 may set the concept scale index to 1.

  In the case of FIG. 7, the conceptual scale index Ibj is shown in the rightmost column of FIG. 9, and in the case of FIG. 8, the conceptual scale index Ibj is in the rightmost column of FIG. Calculated as shown.

  Alternatively, the document analysis system 300 may extract both the word existing immediately before the ambiguous word Waj and the word having a dependency relationship with the word existing immediately before the ambiguous word Waj as the estimated key word Wbj. . In this case, the document analysis system 300 uses a function such as an arithmetic mean or a geometric mean that maintains a monotonically increasing relationship with a plurality of conceptual scale indices Ibj, and sets each estimated key word Wbj related to the same ambiguous word Waj. A value obtained by aggregating the corresponding conceptual scale index Ibj may be used as the final conceptual scale index Ibj.

  Next, the document analysis system 300 divides the conceptual scale index Ibj corresponding to each estimated ambiguous word Waj by the maximum value Ibmax of the conceptual scale index Ib in the document, and normalizes each estimated key word Wbj. It is calculated as an ambiguity index Iaj for each combination of ambiguous words Waj. The document analysis system 300 sets the sentence number of the ambiguity index Iaj and the ambiguous word Waj as the ambiguity information Aj of the ambiguous word Waj. Further, a value obtained by totaling the concept scale index Ibj for the entire document is calculated as ambiguity information Ad representing the quality of the document D.

  For example, in the case of FIG. 9, the ambiguity information Aj is calculated as shown in FIG. FIG. 11 shows that in the general word “document”, the “document etc.” part has a wide estimated range of the concept of “etc.” and is highly ambiguous. In the case of FIG. 10, the ambiguity information Aj is calculated as shown in FIG. FIG. 12 shows that the part of “the output etc. is the same as that of the old system” is highly ambiguous because it is difficult to limit the concept of “etc.” in the highly versatile word “same”.

  Based on the ambiguity information Aj, the document analysis system 300 colors and clearly indicates the combination part of each estimated key word Wbj and the ambiguous word Waj in the document D, such as “document etc.” and “output is the same as the old system”. Thus, the ambiguous sentence to be corrected is displayed in an easy-to-understand manner. Further, the document analysis system 300 aggregates the ambiguity information Ad of the combination of the estimated key word Wbj and the ambiguous word Waj in the entire document D and the chapter unit of the table of contents based on the ambiguity information Ad, and displays a table or graph. It may be output in the format. With the output, the document analysis system 300 provides metrics (evaluation scale) indicating the quality of the document D and information for determining an ambiguous chapter to be corrected.

  As mentioned above, although this invention was demonstrated with reference to each embodiment and an Example, this invention is not limited to the above embodiment and Example. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

  According to the present invention, it is possible to preferentially correct the ambiguity of various documents exchanged in operations such as requirement definition in software and system development from a particularly problematic place. This leads to the efficiency of document creation and document review. In addition, the situation in which different interpretations occur among multiple readers is reduced, and it can be applied to applications related to the efficiency of system development such as reduction of rework and improvement of customer satisfaction.

1 Input Device 2 Data Processing Device 3 Output Device 4 Storage Device 5 ROM
6 RAM
7 CPU
DESCRIPTION OF SYMBOLS 10 Document input part 20 Word analysis part 30, 35 Ambiguous word search part 40, 45 Estimated key word extraction part 50, 55 Concept scale estimation part 60, 65 Ambiguity calculation part 70 Ambiguity output part 100, 200 Document analysis system 110 Ambiguity Word database 120 Concept database 300 Document analysis system 400 Internet server 500 PC terminal 600 Input unit 700 Output unit

Claims (10)

Based on the word information of each word used in the sentences that make up the document input from the outside, search for ambiguous words that are words that have the effect of causing the reader to infer the omitted concepts from the context. When there is an ambiguous word, an ambiguous word search unit that extracts the ambiguous word,
Based on an estimated key word extraction rule for extracting an estimated key word that is a word that becomes a clue when a reader estimates a concept omitted by each ambiguous word extracted by the ambiguous word extraction unit, An estimated key word extraction unit that extracts each estimated key word for an ambiguous word;
A concept scale estimation unit for estimating a concept scale index, which is a concept scale of the estimated key word extracted by the estimated key word extraction unit, based on a concept size estimation rule for estimating a concept size of the estimated key word; ,
Based on the conceptual scale index, an ambiguity calculation unit that calculates ambiguity information for each combination of the ambiguous word and the estimated key word;
Document analysis system including. A document input unit that accepts input of a document;
A word analysis unit that extracts word information, which is information about words used in sentences constituting the document, from the document;
An ambiguous word database that stores ambiguous words that are words that have the effect of causing the reader to infer the omitted concepts from the context;
A concept database that stores general concept information of words;
An ambiguity output unit that outputs ambiguity information;
Further including
The ambiguous word search unit extracts the ambiguous word based on the word information extracted by the word analysis unit and the ambiguous word database,
The conceptual scale estimation unit estimates the conceptual scale index based on the conceptual database and a predetermined conceptual scale estimation rule;
The document analysis system according to claim 1. The predetermined estimated key word extraction rule in the estimated key word extraction unit is a rule for extracting a word existing immediately before the ambiguous word or a word modified by the ambiguous word as the estimated key word.
The document analysis system according to claim 1 or 2. The word analysis unit performs morphological analysis and syntax analysis on the sentences constituting the document, thereby extracting the part of speech of each word and the dependency relationship between the words as the word information,
The predetermined estimated key word extraction rule in the estimated key word extraction unit is a rule for extracting, as an estimated key word, a subject or predicate of a sentence including an ambiguous word or a word having a dependency relationship with a word existing immediately before the ambiguous word Is,
The document analysis system according to claim 1. The concept database is a thesaurus that stores synonym relations, synonym relations, upper relations and lower relations or partial relations and whole relations between words as general concept information.
The document analysis system of any one of Claims 1-4. The predetermined conceptual scale estimation rule in the conceptual scale estimation unit estimates a value that is in a monotonically increasing relationship with the height of the hierarchy as the conceptual scale index using the highest conceptual classification hierarchy to which each estimated key word belongs. Is a rule,
The document analysis system according to claim 5. The predetermined concept scale estimation rule in the concept scale estimation unit is a rule that uses the total number of words registered in the subclass of the concept class to which each estimated key word belongs as a concept scale index.
The document analysis system according to claim 5. The estimated key word extraction unit extracts a plurality of estimated key words for each ambiguous word by having a plurality of estimated key word extraction rules,
The predetermined conceptual scale estimation rule in the conceptual scale estimation unit is an estimated key word related to the same ambiguous word with a function that maintains a monotonically increasing relationship for each of the conceptual scale indices calculated for each of the extracted extracted estimated key words. This is a rule that extracts a value that aggregates the conceptual scale index of as a final conceptual scale index.
The document analysis system according to claim 1. Based on the word information of each word used in the sentences that make up the document input from the outside, search for ambiguous words that are words that have the effect of causing the reader to infer the omitted concepts from the context. If there is an ambiguous word, extract the ambiguous word,
Based on an estimated key word extraction rule for extracting an estimated key word, which is a word that becomes a clue when a reader estimates a concept omitted by each extracted ambiguous word, Extract the estimated key words,
Based on the concept size estimation rule for estimating the concept size of the estimated key word, a concept size index that is the size of the extracted estimated key word concept is estimated,
Based on the conceptual scale index, calculate ambiguity information for each combination of each ambiguous word and estimated key word,
Document analysis method. Based on the word information of each word used in the sentences that make up the document input from the outside, search for ambiguous words that are words that have the effect of causing the reader to infer the omitted concepts from the context. If there is an ambiguous word, extract the ambiguous word,
Based on an estimated key word extraction rule for extracting an estimated key word, which is a word that becomes a clue when a reader estimates a concept omitted by each extracted ambiguous word, Extract the estimated key words,
Based on the concept size estimation rule for estimating the concept size of the estimated key word, a concept size index that is the size of the extracted estimated key word concept is estimated,
Based on the conceptual scale index, calculate ambiguity information for each combination of each ambiguous word and estimated key word,
A program that causes a computer to execute processing.

Images