JP2011227742A - Contrast display data generation device or contrast display data generation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve micro-analysis under the consideration of macro-analysis.SOLUTION: When document data are given, a classification means 5 classifies it into an entity-categorized microscopic viewpoint group or a macroscopic viewpoint group based on an entity specification word. A similarity-categorized group representative keyword determination means 8 classifies it into a similarity-categorized group, and determines a similarity-categorized group representative keyword in each similarity-categorized group. An entity-categorized microscopic viewpoint group representative keyword determination means 7 determines an entity-categorized microscopic viewpoint group representative keyword. A similarity determination means 9 determines similarity from those representative keywords. A generation means 10 determines the combination of extraction from the similarity, and generates relevant display data by arranging the document specification data of the extracted entity-categorized microscopic viewpoint group and the document specification data of the extracted similarity-categorized group on the same time axis for every group.

Description

The present invention relates to a contrast display data generation apparatus, and more particularly to mutual use of macro analysis and micro analysis.

  In Patent Document 1, time-series relationships and events are collected by data mining from various forms of text such as news, blogs, industry reports, and industry papers (magazines) that may include references to various relationships. An entity relationship mining device for extraction is disclosed.

  With such an apparatus, it is possible to analyze a so-called microenvironment such as a trend of other companies in the industry / field directly related to the company of interest.

JP 2005-182528

  However, the above-mentioned Patent Document 1 has a problem that it is difficult to perform analysis in consideration of so-called macro environment such as legal regulations, interest rate trends, and environmental problems that have a great influence on the strategy formulation of a company.

  An object of the present invention is to provide a contrast display data generation apparatus that solves the above-described problems and can present a macro analysis and a micro analysis in association with each other.

  (1) In the comparison display data generating device according to the present invention, A) subject specifying word storage means for storing one or more subject specifying words for specifying the subject, B) text and creation time data are associated with each other. When document data is given, it is determined whether or not the one or more subject specifying words exist in the body of each document data, and if the one or more subject specifying words exist in the body, Classifying means for classifying each subject specified by the subject specifying word into a microscopic viewpoint group for each subject, and when the subject specifying word does not exist, classifying means into a macroscopic viewpoint group, C) A classification result storage means for storing; D) one or more keywords for each document are determined for each document based on the word appearance degree in the body of the document belonging to the macroscopic viewpoint group; Classifying matching documents into groups by similarity, and group representative keywords determining means by similarity for determining one or more groups representative keywords by similarity in each group by similarity, E) designated by the operator, or For subject-specific words predetermined as extraction targets, by subject-specific microscopic viewpoint group based on the degree of appearance in the text of the document belonging to the subject-specific microscopic viewpoint group specified by the subject-specific word A subject-specific microscopic viewpoint group representative keyword determining means for determining one or more microscopic viewpoint group representative keywords; F) each subject-specific microscopic viewpoint group representative keyword and each similarity-specific group document keyword G) similarity determination means for determining the similarity between the subject-specific microscopic viewpoint group and the similarity-specific group; and G) the similarity between the subject-specific microscopic viewpoint group and the similarity-specific group is predetermined. Than the threshold of The document creation data extracted from the subject-specific microscopic viewpoint group and the document data identification data of the document belonging to the similarity-specific group are created on the same time axis for each group. Generating means for generating related display data arranged in chronological order.

  Therefore, it is possible to generate comparison data in which document specifying data belonging to the subject-specific microscopic viewpoint group and document specifying data belonging to the subject-specific microscopic viewpoint group are arranged on the same time axis.

  (2) The comparison display data generating apparatus according to the present invention is provided with A) subject specifying word storage means for storing a subject specifying word for specifying the subject, and B) document data associated with creation time data. Determining whether or not the subject specifying word exists for each document data, and if the subject specifying word exists, classify each subject specified by the subject specifying word into a subject-specific microscopic viewpoint group, Classifying means for classifying into a macroscopic viewpoint group when the subject specifying word does not exist, C) Classification result storing means for storing the classification result, and D) Documents belonging to the macroscopic viewpoint group for each document. Another keyword is determined, classified into groups by similarity based on the determined keywords by document, and group representative key by similarity to determine a group representative keyword by similarity in each group by similarity Word determination means, E) For a subject-specific word designated by the operator or predetermined as an extraction target, a representative keyword is assigned to each subject-specific microscopic viewpoint group for each subject-specific microscopic viewpoint group. A subject-specific microscopic viewpoint group representative keyword determining means; F) a representative keyword for each subject-specific microscopic viewpoint group determined by the subject-specific microscopic viewpoint group representative keyword determining means; Similarity determination means for determining the similarity between the group keyword and G) the similarity between the subject-specific microscopic viewpoint group determined by the similarity determination means and the similarity-specific group group is higher than a predetermined threshold A combination is extracted, and the document creation data extracted from the subject-specific microscopic viewpoint group and the document data identification data of the document belonging to the similarity-specific group are created on the same time axis for each group. And a generating means for generating a related display data arranged in time series order.

  Therefore, it is possible to generate comparison data in which document specifying data belonging to the subject-specific microscopic viewpoint group and document specifying data belonging to the subject-specific microscopic viewpoint group are arranged on the same time axis.

  (3) In the comparison display data generating device according to the present invention, the subject-specific microscopic viewpoint group representative keyword determining means specifies the subject other than the subject-specific word designated by the operator or predetermined as the extraction target. For the word, the representative keyword is determined for each subject-specific microscopic viewpoint group for the documents belonging to each subject-specific microscopic viewpoint group, and the generation means is specified by the operator or predetermined as an extraction target. Among subject-specific microscopic viewpoint groups specified by subject-specific words other than subject-specific words, subject-specific microscopic viewpoint groups that are combinations higher than the predetermined threshold are also related on the same time axis. indicate.

  Accordingly, the subject-specific microscopic viewpoint group specified by the subject specifying word other than the subject specifying word specified by the operator or predetermined as the extraction target can be displayed in a related manner. As a result, a macroscopic viewpoint analysis based on the subject-specific microscopic analysis can be performed on the subject-specific microscopic viewpoint group.

  (4) A comparison display data generation program according to the present invention is a comparison display data generation program for causing a computer to function as the following means. A) Subject specifying word storage means for storing a subject specifying word for specifying the subject, B) If document data associated with creation time data is given, whether or not the subject specifying word exists for each document data If the subject specifying word exists, classify it into a subject-specific microscopic viewpoint group for each subject specified by the subject specifying word, and macroscopic viewpoint group if the subject specifying word does not exist C) Classification result storage means for storing the classification result, D) For each document belonging to the macroscopic viewpoint group, a document-specific keyword is determined for each document, and based on the determined document-specific keywords A group representative keyword deciding means by similarity degree for determining a group representative keyword by similarity degree in each similarity degree group, and E) Is a subject-specific microscopic viewpoint group representative keyword determination that determines a representative keyword for each subject-specific microscopic viewpoint group for a subject belonging to each subject-specific microscopic viewpoint group for a subject-specific word predetermined as an extraction target Means F) a degree of similarity for determining the degree of similarity between the representative keyword for each subject-specific microscopic viewpoint group determined by the subject-specific micro-viewpoint group representative keyword determining means and the keyword for each similarity degree group Determining means, G) extracting a combination in which the similarity between the subject-specific microscopic viewpoint group determined by the similarity determining means and the similarity-specific group group is higher than a predetermined threshold, and extracting the subject-specific microscopic Generation of related display data in which document specific data of document data belonging to a general viewpoint group and document data specific data of a document belonging to a group classified by similarity are arranged for each group in the time series of the creation time on the same time axis Forming means.

  Therefore, it is possible to generate comparison data in which document specifying data belonging to the subject-specific microscopic viewpoint group and document specifying data belonging to the subject-specific microscopic viewpoint group are arranged on the same time axis.

  (5) In the comparison display data generating method according to the present invention, a computer stores a subject specifying word for specifying a subject, and the computer executes the following steps. A) When document data associated with creation time data is given, it is determined whether or not the subject specifying word exists for each document data. If the subject specifying word exists, the subject specifying word is specified. A subject-specific microscopic viewpoint group for each subject to be classified, and when the subject-specific word does not exist, classify into a macroscopic viewpoint group, and store the classification result; B) the macroscopic viewpoint group Determining for each document a document keyword for each document belonging to the document, classifying the keyword into a group by similarity based on the determined document keyword, and determining a group representative keyword by similarity in each similarity group C) Subject-specific words designated by the operator or predetermined as extraction targets, and documents belonging to each subject-specific microscopic viewpoint group, A step of determining a representative keyword for each point group; D) a representative keyword for each subject-specific microscopic viewpoint group determined by the subject-specific microscopic viewpoint group representative keyword determining means; E) extracting a combination having a similarity between the subject-specific microscopic viewpoint group and the similarity group group determined by the similarity determination means being higher than a predetermined threshold Related display in which document specific data of document data belonging to a group of microscopic viewpoints by subject and document data specific data of documents belonging to a group by similarity are arranged for each group in the order of time series of the creation time on the same time axis Generating data.

    Therefore, it is possible to generate comparison data in which document specifying data belonging to the subject-specific microscopic viewpoint group and document specifying data belonging to the subject-specific microscopic viewpoint group are arranged on the same time axis.

It is a functional block diagram of contrast display data generation display device 1. It is a figure which shows an example of the hardware constitution which implement | achieved contrast display data production | generation display apparatus 1 using CPU. It is a figure which shows the data structure of the document memorize | stored in the document memory | storage part 26d. It is a figure which shows the flowchart of the main program 26m. It is a detailed flowchart of step S3 of FIG. It is a figure which shows the data structure of a classification result. It is a figure which shows the group representative word. It is a detailed flowchart of step S5 of FIG. It is a figure which shows a group related result. It is a display example of contrast display data.

  Embodiments of the present invention will be described below with reference to the drawings.

(1. Functional block diagram)
FIG. 1 shows a functional block diagram of a comparison display data generation / display apparatus 1 according to an embodiment of the present invention.

  The comparison display data generation / display apparatus 1 includes a document storage unit 3, a subject identification word storage unit 4, a classification unit 5, a classification result storage unit 6, a similarity-specific group representative keyword determination unit 8, and a subject-specific microscopic viewpoint group representative keyword. A determination unit 7, a similarity determination unit 9, a generation unit 10, and a display unit 11 are provided.

  The document storage means 3 stores a plurality of document data in which the text and the creation time data are associated with each other. The subject specifying word storage unit 4 stores the subject specifying word 1 or two or more for specifying the subject. When the document data in which the text and the creation time data are associated with each other is given from the document storage means 3, the classification means 5 stores the one or two or more stored in the subject specifying word storage means 4 in the text of each document data. If the subject identification word exists, the subject identification word is classified into a subject-specific microscopic viewpoint group for each subject identified by the subject identification word. If it does not exist, classify it into a macroscopic viewpoint group. The classification result storage means 6 stores the classification result.

  The similarity-specific group representative keyword determining means 8 determines one or more document-specific keywords for each document based on the word appearance degree in the body of each document for the documents belonging to the macroscopic viewpoint group. Documents that match different keywords are classified into groups according to similarities, and one or two or more group representative keywords according to similarities in each similarity degree group are determined. The subject-specific microscopic viewpoint group representative keyword determining means 7 belongs to the subject-specific microscopic viewpoint group specified by the subject specifying word for the subject specifying word designated by the operator or predetermined as the extraction target. Based on the degree of appearance in the document, one or more subject-specific microscopic viewpoint group representative keywords of the subject-specific microscopic viewpoint group are determined. Similarity determination means 9 determines the similarity between the subject-specific microscopic viewpoint group and the similarity-specific group based on the degree of coincidence between the subject-specific microscopic viewpoint group representative keyword and the similarity-specific group-specific keyword. Judging. The generation means 10 extracts a combination of the subject-specific microscopic viewpoint group and the similarity-specific group having a similarity higher than a predetermined threshold, and the document data document belonging to the extracted subject-specific microscopic viewpoint group Related display data in which document data specifying data of documents belonging to specific data and similarity groups is arranged for each group in the time series of the creation time on the same time axis is generated. The display means 11 displays the generated related display data.

  As described above, in the present embodiment, since the following processes 1) to 5) are performed, a document belonging to the subject-specific microscopic viewpoint group and a document belonging to the subject-specific microscopic viewpoint group are processed at the same time. Comparison data arranged on the axis can be generated. 1) It is determined whether or not the subject specifying word exists in the body of each document data, and classified into a subject-specific microscopic viewpoint group and a macroscopic viewpoint group. 2) The documents belonging to this macroscopic viewpoint group are classified into groups by similarity based on the word appearance degree in each document, and a group representative keyword by similarity is determined in each group by similarity. 3) Based on the degree of appearance in the document belonging to the subject-specific microscopic viewpoint group, the subject-specific microscopic viewpoint group representative keyword of the subject-specific microscopic viewpoint group is determined. 4) The similarity between the subject-specific microscopic viewpoint group and the similarity-specific group is determined from the degree of coincidence between the subject-specific microscopic viewpoint group representative keyword and the similarity-specific group-specific keyword. 5) extracting the combination of the subject-specific microscopic viewpoint group and the similarity-specific group having a similarity higher than a predetermined threshold, and specifying the document specifying data of the document data belonging to the extracted subject-specific microscopic viewpoint group; Related display data in which document data specifying data of documents belonging to groups by similarity is arranged for each group in the time series of the creation time on the same time axis is generated.

(2. Hardware configuration)
The hardware configuration of the comparison display data generation / display apparatus 1 shown in FIG. 1 will be described with reference to FIG. FIG. 1 shows an example of a hardware configuration in which the comparison display data generation / display apparatus 1 is configured using a CPU.

  The comparison display data generation / display apparatus 1 includes a CPU 23, a memory 27, a hard disk 26, a monitor 30, an optical drive 25, an input device 28 (mouse 28 a, keyboard 28 b), and a bus line 29. The CPU 23 controls each unit via the bus line 29 according to each program stored in the hard disk 26.

  The hard disk 26 stores an operating system program 26o (hereinafter abbreviated as OS) and a main program 26m, and stores various data. In this embodiment, LINUX (registered trademark or trademark) is adopted as the operating system program (OS) 26o, but the present invention is not limited to this.

  Each of the above programs is read from the CD-ROM 25a storing the program via the optical drive 25 and installed in the hard disk 26. In addition to the CD-ROM, a program such as a flexible disk (FD) or an IC card may be installed on a hard disk from a computer-readable recording medium. Furthermore, it may be downloaded using a communication line.

  In the present embodiment, the program stored in the CD-ROM is indirectly executed by the computer by installing the program from the CD-ROM to the hard disk 26. However, the present invention is not limited to this, and the program stored in the CD-ROM may be directly executed from the optical drive 25. Note that programs that can be executed by a computer are not only programs that can be directly executed as they are installed, but also programs that need to be converted to other forms once (for example, those that have been compressed) In addition, those that can be executed in combination with other module parts are also included.

  The data structure of various data stored in the hard disk 26 will be described. A document as shown in FIG. 3 is stored in the document storage unit 26d. Each document has a document ID 301, a date 302, and a body 303. A document ID 301 is an identifier for uniquely identifying a document. Date 302 is the date and time when the document was created. A body 303 is text data representing the contents of the document. For example, the document with the document ID “D001” indicates that the creation date and time is “2003/03/30 016: 43: 00” and the text is “Company X announces portable product OO ...”.

  The company name extraction dictionary 26k stores a plurality of company names to be subjected to micro analysis. In the present embodiment, five company names “Company X”, “Company A”, “Company B”, “Company C”, and “Company D” are stored. The macro classification affiliation determination dictionary 26md is used to determine which of the sub-category categories of the macro analysis is based on the text data described in the body of each document for a document that has not been subjected to micro analysis described later. The classification data is stored. In the present embodiment, a representative keyword is determined for each document, and “politics”, “economics”, “society”, and “technology”, which are classification axes in macro environment analysis, are used for the representative keywords using a concept dictionary. Which of the four categories is determined. For example, WordNet (trademark) (http://wordnet.princeton.edu/) may be adopted as the concept dictionary.

  The data structure of data stored in the classification result storage unit 26r, the group-specific representative word storage unit 26g, and the inter-group relation storage unit 26gk will be described later.

(3. Processing by the main program)
Processing based on the main program 26m will be described with reference to the flowchart of FIG.

  The CPU 23 performs document classification processing (step S1 in FIG. 4). A detailed flowchart of the document classification process is shown in FIG. Hereinafter, a case where the document shown in FIG. 3 is classified as a classification target will be described as an example.

  The CPU 23 initializes the target document number n (step S11). The CPU 23 determines whether or not the document with the object document number n includes a company name (step S13). In this case, since n = 1, the CPU 23 determines whether or not the first document ID “D001” includes the company name in the text. In this case, as described above, “Company X”, “Company A”, “Company B”, “Company C”, and “Company D” are stored in the company name extraction dictionary 26k. Therefore, the CPU 23 determines that the document ID “D001” includes the company name in the body, and sets the classification axis label “Company X” and the type “Company” for the document ID “D001”.

  The CPU 23 stores the classification axis label “Company X” and the type “Company” as the classification result in the classification result storage unit for the document ID “D001” determined in Step S15 (Step S19). The CPU 23 determines whether or not all the processes have been completed (step S21 in FIG. 5). In this case, since the processing has not been completed, the target document number n is incremented.

  The CPU 23 determines whether or not the document with the object document number n includes a company name (step S13). In this case, since n = 2, the CPU 23 determines whether or not the company name is included in the text for the second document ID “D002”. In this case, the CPU 23 determines that the document ID “D002” includes the company name in the body, and sets the classification axis label “X company” and the type “company” for the document ID “D002”.

  The CPU 23 stores the classification axis label “Company X” and the type “Company” as the classification result in the classification result storage unit for the document ID “D002” determined in Step S15 (Step S19).

  Thereafter, the processing from step S23 to step S13 is repeated. If n is incremented as a result of incrementing n in step S23, the eleventh document ID “D011” is determined in step S13. In this case, since the company name is not included in the text, the process proceeds to step S17, where the representative keyword is extracted as an important word for the document ID “D011”, and the macro analysis category determination process is performed (step S17). The representative keyword determination process may be performed by extracting an important word from the body text. For the important word, a keyword is extracted from the text of the body using a morphological analysis engine or the like, and the keyword with the highest frequency may be selected. Also, the structure of the text may be analyzed, and keywords such as the title and the first line of the body that are likely to contain important words may be weighted. In addition, a general tf-idf method or the like may be employed.

  Here, the “mobile site regulation law” is specified as a representative keyword. It is assumed that the representative keyword is determined by the concept dictionary to correspond to the classification axis “politics”.

  For the document ID “D011” determined in step S17, the CPU 23 stores the classification axis label “mobile site regulation law” and the type “politics” as the classification result in the classification result storage unit (step S19).

  Hereinafter, when the processes of steps S13 to S19 are performed for all documents, the document classification process is terminated.

  FIG. 6 shows the classification results stored in the classification result storage unit 26r. In this embodiment, when storing in the classification result storage unit 26r, the CPU 23 assigns and stores the same classification group ID when the value of the classification axis label already exists.

  Next, the CPU 23 performs group-specific representative word determination processing (step S3 in FIG. 4). The group-specific representative word is obtained by performing morphological analysis on all documents belonging to the same group, and by using the tf-idf method or the like, the word having an evaluation value exceeding the threshold is determined by 1 or 2 The determination is as described above. In the present embodiment, it is assumed that group-specific representative words as shown in FIG. 7 are generated.

  Next, the CPU 23 performs relevance determination processing (step S5 in FIG. 4). A detailed flowchart of the relevance determination process is shown in FIG. The CPU 23 acquires a list of classification groups (step S31). The CPU 23 initializes the target group number m (step S33). The CPU 23 extracts another similar group for the target group number m-th group (step S35). In this case, since m = 1, the CPU 23 reads the keywords “Company X”, “mobile”, and “product XX” of the first classification group ID “G001”, and similar to the keywords in other groups Judge the degree. Accordingly, the similarity between groups of the classification group IDs “G001”, “G002”, “G003”,... Is obtained.

  The CPU 23 stores the result in the inter-group relation storage unit 26gk (Step S37). By performing this process for all groups, the degree of association between groups can be obtained. Since the degree of association with another classification group is calculated around the classification group ID “G001”, the degree of association may already be obtained. In that case, the calculation may be omitted.

  FIG. 9 shows the results stored in the inter-group relation storage unit 26gk. In this case, the classification group G001 is similar to the classification group G002, the classification group G001 is similar to the classification group G003, the classification group G001 is similar to the classification group G004, and the classification group G004 is similar to the classification group G005. Relevant results have been obtained.

  CPU23 produces | generates the display data which display the obtained degree of association between groups, and displays it on the monitor 30 (FIG. 4 step S7). In this embodiment, as shown in FIG. 10, the display data to be displayed is generated so that the micro analysis result and the macro analysis result are classified into groups on the same time axis and the creation date and time of each document can be compared. I tried to do it. Specifically, classification groups “G001”, “G002”, and “G003” of the classification axis type “company” are arranged in the micro analysis result display area 510, respectively. In each classification group, document IDs specifying documents belonging to the classification group are arranged in time series. For example, the classification group “G001” is displayed so that it can be seen that the documents D001, D002, and D003 are documents of 2003, 2005, and 2007, respectively.

  In the macro analysis result display area 520, the classification axis type “politics” G004, G005 is displayed. In the group of classification group IDs “G004” and “G005”, document IDs that specify documents belonging to the classification group are arranged in time series. For example, the classification group “G004” is displayed so that it can be understood that the documents D011 and D012 are documents of 2005 and 2009, respectively.

  Therefore, the operator can perform a microeconomic analysis with reference to the macroeconomic analysis result with reference to the display. This provides a business analysis environment apparatus and method for supporting business environment analysis, particularly the external environment surrounding the company based on texts such as news articles and industry reports, and supporting marketing activities or strategy planning / development. be able to.

  In this case, not only the classification groups G002, G003, and G004 related to the classification group G001 but also the classification group G005 similar to the classification group G004 is displayed. Therefore, an analysis that is not assumed by the operator is possible.

  In this case, since there is no corresponding classification group in “economy”, “society”, and “technology” in the macro analysis result display area 520, these are blank.

(4. Other embodiments)
In the present embodiment, the comparison data generation / display apparatus is configured, but the comparison data generation / display apparatus can be grasped. Further, the document data may be acquired from another computer.

  In the present embodiment, the case has been described in which the macro data is further known in four categories (politics, economy, society, technology), but the number of categories may be arbitrary.

  In the present embodiment, all companies stored in the company name extraction dictionary 26k are set as display targets. However, the display target may be further specified by the operator, or the display target may be displayed. May be determined in advance.

  Furthermore, the comparison display data generation display device may select a similar company as a candidate. As a selection method, for a subject belonging to a subject-specific word other than a subject-specific word specified by an operator for a predetermined subject or predetermined as an extraction target, Representative keywords are determined for each group of microscopic viewpoints. Then, among the subject-specific microscopic viewpoint groups specified by the subject specifying word other than the subject specifying word specified by the operator or predetermined as the extraction target, the subject-specific combination that is higher than the predetermined threshold The microscopic viewpoint group may be displayed in association on the same time axis. As a result, companies other than those desired by the operator can be displayed in correspondence with the micro analysis result display area 510 and the macro analysis result display area 520.

  Specifically, assuming that “Company X”, “Company A”, and “Company B” have been specified as micro analysis targets, the operator has data of macro analysis targets related to these and a high publication. In the case of “Company D”, the micro analysis results of “Company C” and “Company D” can be displayed together.

  In the above embodiment, the representative word for each group is determined from all the documents belonging to each group, a vector is generated based on the appearance frequency of each representative word for each group, and the inner product value is used as the similarity between documents. Defined and acquired high similarity. However, the present invention is not limited to this, and the degree of similarity between the group-specific types of documents may be determined by combining the body texts of all the documents belonging to each group as group-specific types of documents.

  In the above embodiment, the CPU 23 is used to realize the function shown in FIG. 1, and this is realized by software. However, some or all of them may be realized by hardware such as a logic circuit. In addition, you may make it make an operating system (OS) process a part of program.

Claims (5)

Subject specifying word storage means for storing one or more subject specifying words for specifying the subject,
When the document data in which the body text and the creation time data are associated with each other is given, it is determined whether or not the one or more subject specifying words exist in the body of each document data, and the one or two or more body identification words are included in the body. If there is a subject-specific word, classify the subject-specific microscopic viewpoint group for each subject specified by the subject-specific word, and if there is no subject-specific word, classify it as a macroscopic viewpoint group Classification means,
Classification result storage means for storing the classification result;
Based on the word appearance degree in the body text of the document belonging to the macroscopic viewpoint group, one or two or more document keywords are determined for each document, and the documents matching each document keyword are classified into the similarity group. , Group representative keyword determination means by similarity for determining one or more group representative keywords by similarity in each group by similarity,
For subject-specific words designated by the operator or predetermined as extraction targets, the subject-specific words are determined based on the appearance level of the document belonging to the subject-specific microscopic viewpoint group specified by the subject-specific words. A subject-specific microscopic viewpoint group representative keyword determining means for determining one or more subject-specific microscopic viewpoint group representative keywords of the visual viewpoint group;
Similarity determination means for determining the similarity between the subject-specific microscopic viewpoint group and the similarity-specific group from the degree of match between the subject-specific microscopic viewpoint group representative keyword and the similarity-specific group-specific keyword ,
The combination of the subject-specific microscopic viewpoint group and the similarity-specific group is extracted with a combination higher than a predetermined threshold, and the document specifying data and similarity of the document data belonging to the extracted subject-specific microscopic viewpoint group Generating means for generating related display data in which document data specifying data of documents belonging to different groups are arranged for each group in time series of the creation time on the same time axis;
A comparison display data generation device. A subject specifying word storage means for storing a subject specifying word for specifying the subject,
When document data associated with creation time data is given, it is determined whether or not the subject specifying word exists for each document data. If the subject specifying word exists, the subject specifying word is specified by the subject specifying word. Classifying means for classifying each subject according to a microscopic viewpoint group for each subject, and when there is no subject specific word, classifying into a macroscopic viewpoint group,
Classification result storage means for storing the classification result;
For each document belonging to the macroscopic viewpoint group, a keyword for each document is determined for each document, classified into a group by similarity based on the determined keyword for each document, and a group representative by similarity in each group by similarity Group representative keyword determination means by similarity to determine keywords,
Subject-specific micros to determine representative keywords for each subject-specific microscopic viewpoint group for documents belonging to each subject-specific microscopic viewpoint group for subject-specific words specified by the operator or predetermined as extraction targets Representative viewpoint group representative keyword decision means,
Similarity determination means for determining the similarity between the representative keyword for each subject microscopic viewpoint group determined by the subject-specific microscopic viewpoint group representative keyword determination means and the keyword for each similarity degree group;
A combination in which the similarity between the subject-specific microscopic viewpoint group determined by the similarity determination means and the similarity-specific group group is higher than a predetermined threshold is extracted and belongs to the extracted subject-specific microscopic viewpoint group Generating means for generating related display data in which document specific data of document data and document data specific data of a document belonging to a group classified by similarity are arranged for each group in time series on the same time axis;
A comparison display data generation device. In the comparison display data generation device according to claim 2,
The subject-specific microscopic viewpoint group representative keyword determining means belongs to the subject-specific microscopic viewpoint group even for subject-specific words other than the subject-specific words specified by the operator or predetermined as extraction targets. For a document, determine one or more representative keywords for each group of microscopic viewpoints by subject,
The generation means includes a combination higher than the predetermined threshold among the subject-specific microscopic viewpoint groups specified by the subject specifying word other than the subject specifying word specified by the operator or predetermined as the extraction target. The subject-specific microscopic viewpoint group is related and displayed on the same time axis,
A comparison display data generation device characterized by the above. A comparison display data generation program for causing a computer to function as the following means.
A subject specifying word storage means for storing a subject specifying word for specifying the subject,
When document data associated with creation time data is given, it is determined whether or not the subject specifying word exists for each document data. If the subject specifying word exists, the subject specifying word is specified by the subject specifying word. Classifying means for classifying each subject according to a microscopic viewpoint group for each subject, and when there is no subject specific word, classifying into a macroscopic viewpoint group,
Classification result storage means for storing the classification result;
For each document belonging to the macroscopic viewpoint group, a keyword for each document is determined for each document, classified into a group by similarity based on the determined keyword for each document, and a group representative by similarity in each group by similarity Group representative keyword determination means by similarity to determine keywords,
Subject-specific micros to determine representative keywords for each subject-specific microscopic viewpoint group for documents belonging to each subject-specific microscopic viewpoint group for subject-specific words specified by the operator or predetermined as extraction targets Representative viewpoint group representative keyword decision means,
Similarity determination means for determining the similarity between the representative keyword for each subject microscopic viewpoint group determined by the subject-specific microscopic viewpoint group representative keyword determination means and the keyword for each similarity degree group;
A combination in which the similarity between the subject-specific microscopic viewpoint group determined by the similarity determination means and the similarity-specific group group is higher than a predetermined threshold is extracted and belongs to the extracted subject-specific microscopic viewpoint group Generating means for generating related display data in which document specific data of document data and document data specific data of a document belonging to a group classified by similarity are arranged for each group in the time series of the creation time on the same time axis; A comparison display data generation method in which a subject specifying word for specifying a subject is stored in a computer, and the computer executes the following steps.
When document data associated with creation time data is given, it is determined whether or not the subject specifying word exists for each document data. If the subject specifying word exists, the subject specifying word is specified by the subject specifying word. Classifying each subject into a microscopic viewpoint group for each subject, classifying into a macroscopic viewpoint group if the subject specific word does not exist, and storing the classification result;
For each document belonging to the macroscopic viewpoint group, a keyword for each document is determined for each document, classified into a group by similarity based on the determined keyword for each document, and a group representative by similarity in each group by similarity Determining keywords,
Determining a representative keyword for each subject-specific microscopic viewpoint group for a document belonging to each subject-specific microscopic viewpoint group for a subject-specific word designated by an operator or predetermined as an extraction target;
Determining a similarity between a representative keyword for each subject microscopic viewpoint group determined by the subject-specific microscopic viewpoint group representative keyword determining means and a keyword for each similarity degree group;
A combination in which the similarity between the subject-specific microscopic viewpoint group determined by the similarity determination means and the similarity-specific group group is higher than a predetermined threshold is extracted and belongs to the extracted subject-specific microscopic viewpoint group Generating related display data in which the document specification data of the document data and the document data specification data of the documents belonging to the groups classified by similarity are arranged for each group in the time series of the creation time on the same time axis.

Images