JP2009128949A - Graphic display device and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To suggest the tendency of a whole text unit aggregation by graphically indicating a plurality of representative words in the whole text unit aggregation to be analyzed, and relations among the words. <P>SOLUTION: A word appearance data storage part 150 holds word appearance data which indicates which message includes the word. A frequency calculating part 170 calculates how many messages exist in the object message aggregation, where the respective object words appear. A word specifying part 180 specifies L number of words with the first to L-th frequencies, based on the frequencies calculated by the frequency calculating part 170. A partial aggregation deriving part 160 derives partial aggregations, specifies the corresponding words, and generates node data and link data. A graphic display part 190 acquires the node data and the link data from the partial aggregation deriving part 160, determines a graphic network structure, and performs graphic display. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for graphically displaying words included in a document.

  Analyzing a large number of documents such as questionnaires and complaint documents (emails) that are answered in free format, and extracting the truth inherent in these large numbers of documents. For example, it is desirable to analyze the message in the free answer column and extract requests, thanks, satisfaction, requirements, dissatisfaction, etc., and feed it back to product development. Such an extraction operation is usually performed by a skilled worker taking a lot of time, and there is a problem in terms of cost or immediacy.

  It is desirable to support such work using language processing by a computer.

  Incidentally, Patent Document 1 is related to the present invention. In Patent Document 1, a feature word having a large ratio of the frequency appearing in the search result document with respect to the frequency appearing in the entire search target document is represented by a node, and a link is established between the feature word nodes based on the co-occurrence degree of the feature word The graph display is disclosed. However, it is based on search and does not support the analysis of documents in general situations.

It should be understood that the above-described background art and its problems are only part of the background of the present invention, and the present invention is not limited to the above-mentioned background techniques and problems.
JP-A-10-74210

  The present invention has been made in consideration of the above circumstances, and provides a graph display technique for displaying words in a graph to analyze the words contained in the entire target document and assist in analyzing the entire target document. The purpose is to do.

  According to this invention, in order to achieve the above-mentioned object, the configuration as described in the claims is adopted. Here, prior to describing the invention in detail, supplementary explanations of the claims will be given.

  That is, according to one aspect of the present invention, in order to achieve the above-described object, a graph display device that graphically displays words that appear in a sentence unit set to be analyzed including a plurality of sentence units: Storage means for storing words appearing in the sentence unit in association with each other; frequency calculation means for calculating a frequency in the sentence unit set with reference to the storage means for words appearing in the sentence unit in the sentence unit set; Referring to the frequency calculated by the frequency calculating means, a word specifying means for specifying the uppermost L (an integer greater than or equal to 2) words in the sentence unit set; a word specified by the word specifying means; A subset derivation means for deriving a subset of the sentence unit including the word from the sentence unit set; and a derivation destination from the word corresponding to the subset and the word corresponding to the derivation sentence unit set Part of And display means for displaying a link to a word corresponding to the set provided, said subset derivation means is so as to derive a subset repeated starting from the statement unit set route.

  Typically, the graph display apparatus can be configured by a stand-alone computer system, a computer system group (client / server system) connected to a network, or the like, but is not limited thereto.

  A sentence unit is a single sentence composed of one or a plurality of sentences, and may be referred to as a message below. The sentence unit may be a part of a sentence such as a segment. A word is typically a noun, but is not limited thereto. A simple display can be realized by limiting the word to a specific part of speech, for example a noun.

  In this configuration, a subset is sequentially derived from the sentence unit set, focusing on the local word frequency of the sentence unit set, and the direction and the words corresponding to the derived subset are displayed as a graph. Then, a plurality of typical words in the whole sentence unit set to be analyzed and the relationship between the words can be displayed in a graph to suggest a tendency of the whole sentence unit set.

  The subset may be derived by depth-first search and width-first search. Also, subsets may be derived in descending order of frequency in the derivation source sentence unit set.

  The storage unit may store one or a plurality of attributes for each sentence unit, and narrow down a sentence unit set to be analyzed according to the attribute. The attribute is typically gender and age (age range), and the display may be switched according to the attribute value or displayed in comparison with the attribute value to suggest the feature of the attribute value.

  The word specifying means may prohibit the same word from being specified more than a predetermined number of times. For example, the same word may be specified only once.

  According to another aspect of the present invention, there is provided a graph display device for graphically displaying words appearing in a set of sentence units to be analyzed including a plurality of sentence units: appearing in the sentence unit in association with the sentence unit A frequency calculation for calculating a frequency of co-occurring first and second words that make a pair in a sentence unit in the sentence unit set to be analyzed with reference to the storage means Means; for each pair whose frequency calculated by the frequency calculation means exceeds a threshold, the frequency of the pair is up to the upper L (an integer of 2 or more) of the pairs having the same first word of the pair A word specifying means for specifying the first word and the second word that make up the pair, the first word and the second word specified by the word specifying means, and Display means for displaying a graph consisting of links linking The are to be provided.

  In this configuration, the word can be displayed in a graph based on the relevance of the word, focusing on the co-occurrence frequency of the word in the entire target sentence unit set. By providing a threshold for the co-occurrence frequency and screening, it is possible to prevent the marking of the graph from becoming complicated.

  The present invention can be realized not only as an apparatus or a system but also as a method. Of course, a part of the invention can be configured as software. Of course, software products used to cause a computer to execute such software are also included in the technical scope of the present invention.

  These and other aspects of the invention are set forth in the appended claims and will be described in detail below with reference to examples.

  According to the present invention, a subset is sequentially derived from a sentence unit set by paying attention to the local word frequency of the sentence unit set, and the derived direction and the word corresponding to the derived subset are graphed. A plurality of typical words in the whole sentence unit set to be analyzed and the relationship between the words can be displayed in a graph to indicate a tendency of the whole sentence unit set.

  Examples of the present invention will be described below.

  FIG. 1 shows an overall graph display system 110 according to a first embodiment for explaining the basic configuration of the present invention. In this example, it is realized as software on the computer 200. The software can be installed in the computer 200 using the recording medium 201 by a well-known method or using a communication line. In the figure, a stand-alone configuration is used, but a server device and a client device connected via a network may be used. Each functional block of the graph display system 110 according to the first embodiment is typically realized by cooperation of software and hardware resources of the computer 200.

  For example, the graph display system 110 displays a graph including a word included in a message with a tendency of a free-form message (a sentence unit composed of one or more sentences) included in the answer for analysis of a questionnaire response. Present. A fixed-format answer (choice or fixed word) may be used together.

  In this embodiment, as shown in FIG. 2, the word to be displayed and the relationship (link) between the words are specified. Specifically, the following processing is performed.

(1) The frequency of words included in the message set is obtained.
(2) Create L subsets in descending order of frequency. Hereinafter, L may be referred to as a width.
(3) Process (2) is performed for each subset. Repeat this M times. Hereinafter, M is sometimes referred to as depth.

  In this way, the subset is derived with the width L and the depth M, the word used to create the derived subset is selected as the display target word, and the relationship from the introduction source set to the introduction destination set is introduced. From the word specified from the original set, a link from the introduction destination set to the specified word is generated, and a graph including the specified word and link is generated and displayed.

  1, the graph display system 110 includes a word appearance data storage unit 150, a frequency calculation unit 170, a word identification unit 180, a subset derivation unit 160, a graph display unit 190, and the like.

The word appearance data storage unit 150 holds, for example, word appearance data having a data structure as shown in FIG. In this example, the word appearance data includes a message number, a word ID, and a plurality of attribute values assigned to the message (questionnaire answer), but is not limited thereto. The attributes are sex, age (age range), place of residence, and the like. When N different words are included in one message, N word appearance data are prepared. The word to be prepared may be limited by part of speech such as a noun.
Specific examples of word appearance data are not limited to this, but are as shown in FIG. 5, for example, and include message numbers, dependency relationships that receive predicates, messages (sentences), and message attribute values. ("Fact" column). For example, the data on the top line has a message number of 10419, the predicate (receiving part) is “matching”, and the noun part (relative part) is “skin”. Because the dependency part can be displayed by classifying the particles as “ha”, “ga”, “on”, “ni”, “de”, etc., it is easy to understand the meaning content as a whole simply by referring to the dependency become.

  The frequency calculation unit 170 calculates how many messages each of the target words appear in the target message set. Typically, the frequency calculation unit 170 stores the word appearance data including the word appearance data for each word ID. Although the number is calculated, any method may be adopted as long as it generates a histogram of the number of messages for each word.

  Based on the frequency calculated by the frequency calculation unit 170, the word specifying unit 180 specifies L words having frequencies from the first place to the L place.

The subset derivation unit 160 derives a subset according to the procedure described with reference to FIG. Get the specified word from. The subset deriving unit 160 obtains node data as shown in FIG. 6A from the identified word, and generates link data as shown in FIG. 6B based on the introduction source and introduction destination of the subset. get. The node data includes a node ID, a word ID, a word (character string), the number of messages (number of elements) included in a subset corresponding to the node, but is not limited thereto. The link data includes, but is not limited to, a link ID, a node ID of a source node (parent, derivation source), a node ID of a target node (child, derivation destination), and the like.
The graph display unit 190 acquires node data and link data from the subset derivation unit 160, determines the network structure of the graph, and displays the graph. The network structure can be determined by an arbitrary method, but is not limited thereto, but is not limited to “key graph” (http://www2.kke.co.jp/keygraph) or “GraphViz” (http://i.loverruby.net/ ja / rhg / cd / graphviz.html). A display example of the graph display unit 190 is, for example, as shown in FIG. 9 and links words appearing in the entire target message set with links. Words (nodes), link size, thickness, color, and the like may be varied according to the frequency of messages.

  FIG. 3 shows an example of the operation of this embodiment, and the processing is as follows.

[Step S10]: The entire set of messages is set as a frequency calculation target.
[Step S11]: The word frequency is calculated.
[Step S12]: L or less words with the highest frequency are selected. For example, when searching for a subset with depth priority (backtrack method), the subsets 1 to N are searched for as shown in FIG. Select one word. When searching for a subset with priority in width, the subsets 1 to N are searched as shown in FIG. 6B (example of L = 2).
[Step S13]: Node / link data is generated and stored. As the storage location, the word appearance data storage unit 150 may be used, or any other storage means may be used.
[Step S14]: A subset corresponding to the selected word is set as a frequency calculation target.
[Step S15]: It is determined whether or not the end condition is satisfied. If not satisfied, the process returns to Step S11, and if satisfied, the process proceeds to Step S16. The end condition is when a word has been specified to a specified depth, when an upper limit number of words has been specified, or when there are no more sets containing messages exceeding the lower limit number, but it is not limited to this. .
[Step S16]: The graph display unit 190 generates graph data by a predetermined graph generation method using the node data and link data, and displays the graph.

  8 shows the message frequency for each word of the entire set of messages as a bar graph (FIG. 8 (a), which is further displayed as a word graph (M is set to 1). Then, since the hierarchical processing as shown in FIG. 2 is not performed, it merely shows which words appear frequently.

  FIG. 9 shows a set of messages prepared from an answer to a free-form questionnaire for cosmetics (“the foundation that you currently use / what you like”) using the graph display system 110 of this embodiment. A display example of the generated graph is shown. It is possible to grasp what is being told by following the links of the graph. For example, a group of words is grasped as a partial graph, and the meaning is extracted from each. In the example of FIG. 9, it can be understood that (1) a color suitable for oneself, (2) price / quality, (3) finish, (4) growth, and the like are included as meanings.

  10 and 11 are graphs stratified by age range. FIG. 10 is a graph in which messages of respondents aged 25 to 29 are generated as a whole set based on age attribute values, and FIG. 11 is a graph of messages of respondents aged 40 to 49 years as a whole set. It is a graph. From FIG. 10, it can be understood that the feeling of use is not added and that the skin is sensitive. It can be understood from FIG. 11 that the covering power and lightness are talking about the stain and also the UV effect.

  In this embodiment, it becomes easy to grasp what is a topic based on the arrangement of words distributed on the graph.

  Next, a graph display system 100 according to a second embodiment to which the present invention is specifically applied will be described.

  FIG. 12 shows a graph display system 100 according to a second embodiment of the present invention as a whole. In this example, the graph display system 100 is realized as software on the computer 200. The software can be installed in the computer 200 using the recording medium 201 by a well-known method or using a communication line. In the figure, a stand-alone configuration is used, but a server device and a client device connected via a network may be used. Each functional block of the graph display system 110 according to the second embodiment is typically realized by cooperation of software and hardware resources of the computer 200. In FIG. 12, portions corresponding to those in FIG.

  12, the graph display system 100 includes a document input unit 10, a morpheme analysis unit 11, a syntax analysis unit 12, a syntax analysis result data storage unit 13, a dependency relationship extraction unit 14, a dependency relationship set storage unit 15, and a subset. A derivation unit 16, a frequency calculation unit 17, a word identification unit 18, a graph display unit 19 and the like are included. In this example, a set of dependency relationships is extracted from the document data consisting of a group of sentences and stored in the dependency relationship set storage unit 15. However, the dependency relationship set is acquired from the outside to obtain the dependency relationship. It may be stored in the collective storage unit 15.

  In this embodiment, dependency-related data extracted from the document data (referred to as basic semantic chunks as described below, which are grouped within a common range of receiving parts such as predicates). , Focusing on the dependency relationship, it is possible to search for the dependency word, the received word, and other words, and typically display the comparison of the dependency word and the received word to assist the comprehension of the sentence. .

  The document input unit 10 inputs document data (also called a corpus) such as a group of questionnaires (an answer to a free-form questionnaire, a message) or a group of e-mails, and the document data performs subsequent morphological analysis. Therefore, it may be preprocessed appropriately. The document input unit 10 may be anything as long as it can input document data, and includes, for example, a file system, an external storage device, a communication line, and an I / O device. The document input unit 10 may be a system that receives a message such as a questionnaire or an e-mail, or may be a character recognition device, a voice recognition device, or the like. An example of document data is as shown in FIG. 14, for example. In this example, sentences acquired from cosmetic questionnaire results are managed with sentence numbers assigned.

  The morpheme analyzing unit 11 decomposes a sentence into morphemes by referring to a morpheme analysis dictionary by a known arbitrary morpheme analysis method. The morphological analysis is performed as shown in FIG. 15, for example.

  The syntax analysis unit 12 parses the morpheme analysis result based on the syntax rules by any known syntax analysis method. That is, as shown in FIG. 13, a group of sentence data (corpus) is input by the document input unit 10 (S110). The morpheme analysis unit 11 takes out data of one sentence as a processing target, performs morpheme analysis, and the syntax analysis unit 12 performs syntax analysis based on the morpheme analysis result (S111 to S113). The syntax analysis result is registered in the syntax analysis result data storage unit 13, and the above processing is repeated for all sentences (S114, S115). The parsing result is as shown in FIG. 16, for example, and when this is expressed in a tree structure for easy understanding, it is as shown in FIG.

  The dependency relationship extracting unit 14 applies a dependency relationship extraction rule to the data of the syntax analysis result, extracts a dependency relationship set, and records it in the dependency relationship set storage unit 15. The dependency relationship extraction rule is as shown in FIG. 19, for example. In the figure, “*” is an arbitrary number of sub-trees (elements forming a part of the parse tree). In this example, a continuous dependency relationship is extracted, but the same applies to the dependency relationship. Using the dependency relationship extraction rule, for example, the dependency relationship can be extracted as indicated by an arrow in FIG. In this example, a continuous dependency relationship is shown.

  As shown in FIG. 18, the dependency relationship extraction unit 14 inputs an extraction rule (S120), extracts the corresponding dependency relationship from the data of the syntax analysis result (S121), and the dependency relationship set storage unit 15 (S122).

  The extracted dependency-related data is expressed in a unified form within a range in which receiving parts such as phrases are shared. There are zero or a plurality of engaging portions. Hereinafter, such data is also referred to as a basic semantic chunk. The basic semantic chunk is represented by, for example, a data structure in the fact format of the programming language Prolog, and FIG. 21 shows an example of such a data structure. This data structure consists of two types of data: basic semantic chunks and chunk predicates. In the example of FIG. 21, “1” indicates a sentence number, “Introduction, To Do,” indicates a morpheme sequence of prescriptive phrases arranged in an end form, and “23, 31” indicates an appearance position thereof. “Taro, ha”, “5, 11”, “Hanako, o”, “17, 23”, “Jiro, ni”, “11, 17” are the morpheme sequence of the receiving part and the respective This is a byte offset that indicates the appearance position. “3” indicates the number of dependencies. “Introduction_example sentence” is the name of the corpus.

  The dependency relationship set storage unit 15 stores a dependency relationship set (basic semantic chunk set). The dependency relationship set storage unit 15 corresponds to the word appearance data storage unit 150 of the first embodiment, and the dependency relationship set is the same as that shown in FIG. 5, for example.

  The frequency calculation unit 17, the word identification unit 18, the subset derivation unit 16, and the graph display unit 19 correspond to the frequency calculation unit 170, the word identification unit 180, the subset derivation unit 160, and the graph display unit 190 of the first embodiment. To do.

  That is, the frequency calculation unit 17 calculates how many messages in each target word appear in the target message set, and typically calculates the number of messages including the word for each word. .

  Based on the frequency calculated by the frequency calculation unit 17, the word specifying unit 18 specifies L words whose frequencies are from 1st to Lth.

The subset derivation unit 16 derives the subset by the procedure described above with reference to FIG. 2. At this time, the introduction source set is designated to the frequency calculation unit 17 to specify the word. The specified word is acquired from the unit 18. The subset deriving unit 16 obtains the node data as shown in FIG. 6A from the specified word, and shows the node data shown in FIG. 6B based on the introduction source and the introduction destination of the subset. Get link data like this. The node data includes a node ID, a word ID, a word (character string), the number of messages (number of elements) included in a subset corresponding to the node, but is not limited thereto. The link data includes, but is not limited to, a link ID, a node ID of a source node (parent, derivation source), a node ID of a target node (child, derivation destination), and the like.
The graph display unit 19 acquires node data and link data from the subset derivation unit 16, determines the network structure of the graph, and displays the graph. The network structure can be determined by any method. A display example will be described below. However, the size, thickness, color, and the like of words (nodes) and links may be varied according to the frequency of messages.

  FIG. 22 shows an example of a user interface for setting graph display parameters in this embodiment. In this example, the subset derivation mode is selected from depth priority and width priority using the radio buttons R1 and R2, the fan-out (width L) and the depth M are input to the input forms F1 and F2, and the fan-in (node Specifies the number of links to be input to (the number of subsets that can be selected by specifying the same word). As display options, display / non-display of the entire node and frequency display can be designated by radio buttons R3 and R4. A default value may be provided for the parameter. After setting the graph display parameters, the graph display button B is operated to start the graph display processing. The graph display process is typically executed as shown in FIG.

  FIG. 23 shows a graph when the fan-out 10, depth 2, fan-in 1, width priority, overall node display, and frequency display (number of messages of the target node) are set. In this example, since breadth is given priority, frequently used words are displayed, so it is easy to grasp the topic. However, the relationship between topics (frequently used words) is slightly difficult to grasp.

  FIG. 24 shows a graph when the setting is fan-out 10, depth 4, fan-in 1, depth priority, overall node display, and frequency display. In this example, frequently used words (topics) can be connected in a co-occurrence relationship, and the relationship between topics can be easily grasped. However, when the depth is shortened, even a frequently used word may not be displayed (when searching from the left, the rightmost word in the upper hierarchy).

  FIG. 25 shows a graph when the fan-out 10, depth 3, fan-in 2, width priority, overall node display, and frequency non-display are set. In this example, as the fan-in increases, it becomes easier to grasp the relationship between topics (words), but the graph display becomes complicated.

  FIG. 26 shows a graph when the fan-out 10, depth 3, fan-in 3, width priority, overall node display, and frequency non-display are set. In this example, as the fan-in increases, it becomes easier to grasp the relationship between topics (words), but the graph display becomes more complicated.

  FIG. 27 shows a graph when the setting is fan-out 10, depth 5, fan-in 1, width priority, overall node display, and frequency non-display. In this example, you can follow the relationship of the topic in a wide range by increasing the depth. However, there is a case where it becomes noise when grasping the topic. The frequency (the frequency of messages in the target node, not the frequency of words in the entire message set) may be displayed, or the color may be changed according to the frequency to emphasize important topics.

  FIG. 28 shows a graph when the setting is fan-out 10, depth 3, fan-in 1, width priority, overall node display, and frequency non-display. In this example, a wide variety of topics can be displayed.

  In the above display example, the user may designate a region of interest so as to be surrounded by a chain line in FIG.

  Next, a graph display system 120 according to Embodiment 3 of the present invention will be described.

  The graph display system 120 of this embodiment selectively displays words by “edge frequency priority” in addition to depth priority and width priority. Here, the “edge frequency” refers to the number X of messages including another word W in a set of messages including a predetermined word V in the entire target message set. This is represented by V → W (X). For example, mobile phone → children (100) means that there are 100 messages including the word “child” in the message set including the word “mobile phone”. The original word of → is called the source and the previous word is called the target.

  FIG. 29 shows the graph display system 120 of the third embodiment as a whole, and in this figure, parts corresponding to those in FIG. In this example, it is realized as software on the computer 200. The software can be installed in the computer 200 using the recording medium 201 by a well-known method or using a communication line. In the figure, a stand-alone configuration is used, but a server device and a client device connected via a network may be used. Each functional block of the graph display system 110 according to the first embodiment is typically realized by cooperation of software and hardware resources of the computer 200.

  29, the graph display system 120 includes a word appearance data storage unit 150, an edge frequency calculation unit 171, a word pair identification unit 181, a graph display unit 190, and the like. Although this graph display system 120 is not shown in FIG. 29, it also has the configuration shown in FIG. Of course, you may comprise only the part required for the display of priority of edge frequency. The edge frequency calculation unit 151 refers to the word appearance data storage unit 150 and calculates the edge frequency V → W (X) for two words V and W. The word pair identification unit 181 sorts the word pairs (V, W) with the edge frequency V → W (X), identifies word pairs with a frequency greater than the set frequency, and generates node data and link data. The upper predetermined number of word pairs may be specified. The graph display unit 190 displays a graph based on the node data and link data received from the word pair identification unit 181.

  FIG. 30 illustrates an operation example of the graph display system 120, and details thereof are as follows.

[Step S20]: The entire set of messages is set as a target.
[Step S21]: The edge frequency V → W (X) is calculated.
[Step S22]: Sort word pairs (V, W) by edge frequency V → W (X).
[Step S23]: A word pair (V, W) whose edge frequency exceeds a threshold is specified. You may select by other conditions, for example, the number of word pairs.
[Step S24]: Node data and link data are generated from the specified word pair (V, W).
[Step S25]: A graph including nodes and links is displayed.

  FIG. 31 shows an example of a user interface for setting graph display parameters. In this example, the word search mode (subset derivation mode) is specified with the radio button R4 including the edge frequency priority in addition to the width priority and depth priority, and the edge frequency threshold to be displayed is specified with the input form F4. It is supposed to be.

  FIG. 32 shows a graph when setting is made with edge frequency priority, fan-in 3, edge frequency threshold 10, frequency non-display, and entire node display. Also in this example, by appropriately setting the edge frequency threshold, pruning can be performed and the relationship between topics can be displayed in a non-complex manner. Pruning may be performed by limiting the number of words to be displayed.

  In addition, this invention is determined based on description of a claim, and is not limited to the specific structure of the Example, a subject, and an effect. The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention. For example, a dependency relationship or message including a corresponding word may be displayed as shown in FIG. 5 by clicking the word node in the graph display.

It is a block diagram explaining the structure of Example 1 of this invention. It is a figure which illustrates operation | movement of the said Example 1 typically. It is a flowchart explaining the operation example of the said Example 1. FIG. It is a figure explaining the word appearance data of the said Example 1. FIG. It is a figure explaining the word appearance data of the dependency relation of the above-mentioned Example 1. It is a figure explaining the node data and link data of the said Example 1. FIG. It is a figure explaining the example which derives a subset by depth priority and width priority of the above-mentioned Example 1. It is a figure explaining the conventional display mode. It is a figure explaining the example of a display of the above-mentioned Example 1. FIG. It is a figure explaining the example of a display classified according to the attribute value of the above-mentioned Example 1. It is a figure explaining the other example of a display classified according to the attribute value of the above-mentioned Example 1. It is a block diagram which shows the structure of Example 2 of this invention as a whole. It is a flowchart explaining the example of a syntax analysis result acquisition operation | movement of the said Example 2. FIG. It is a figure explaining the example of the document data of the said Example 2. FIG. It is a figure explaining the example of the morphological analysis of the above-mentioned Example 2. FIG. It is a figure explaining the example of the syntax analysis result of the said Example 2. FIG. It is a figure explaining the tree structure expression of the syntax analysis result of the above-mentioned Example 2. It is a flowchart explaining the operation example of the dependency relationship extraction of the said Example 2. FIG. It is a figure explaining the example of the dependency relationship extraction rule of the said Example 2. FIG. It is a figure explaining the example of the extraction result of the dependency relationship of the said Example 2. FIG. It is a figure explaining the example of the data structure of the dependency relation of the said Example 2. FIG. It is a figure explaining the example of a user interface which sets the parameter of the graph display of the said Example 2. FIG. It is a figure explaining the example of a graph display of the above-mentioned Example 2. It is a figure explaining the other graph display example of the said Example 2. FIG. It is a figure explaining the other graph display example of the said Example 2. FIG. It is a figure explaining the other graph display example of the said Example 2. FIG. It is a figure explaining the other graph display example of the said Example 2. FIG. It is a figure explaining the other graph display example of the said Example 2. FIG. It is a block diagram which shows the structure of Example 3 of this invention as a whole. It is a flowchart explaining the operation example of the said Example 3. FIG. It is a figure explaining the example of a user interface which sets the parameter of the graph display of the said Example 3. FIG. It is a figure explaining the example of a graph display of the above-mentioned Example 3.

Explanation of symbols

110 Graph Display System 150 Word Appearance Data Storage Unit 160 Subset Derivation Unit 170 Frequency Calculation Unit 180 Word Identification Unit 190 Graph Display Unit

Claims (11)

In a graph display device that graphically displays words that appear in a set of sentence units to be analyzed including a plurality of sentence units,
Storage means for storing words appearing in the sentence unit in association with the sentence unit;
Frequency calculation means for calculating the frequency in the sentence unit set with respect to words appearing in the sentence unit in the sentence unit set with reference to the storage means;
Referring to the frequency calculated by the frequency calculating means, a word specifying means for specifying the uppermost L (an integer greater than or equal to 2) words in the sentence unit set;
For each word specified by the word specifying means, a subset derivation means for deriving a subset of sentence units including the word from the sentence unit set;
Display means for displaying a link from the word corresponding to the subset and the word corresponding to the sentence unit set of the derivation source to the word corresponding to the subset of the derivation destination;
The graph display device, wherein the subset derivation means derives a repeated subset starting from a sentence unit set of a root.   The graph display device according to claim 1, wherein the subset derivation unit derives the subset by a depth-first search.   The graph display device according to claim 1, wherein the subset derivation means derives a subset by breadth-first search. In a graph display device that graphically displays words that appear in a set of sentence units to be analyzed including a plurality of sentence units,
Storage means for storing words appearing in the sentence unit in association with the sentence unit;
A frequency calculating means for calculating the frequency with which the first word and the second word that make a pair in the sentence unit in the sentence unit set to be analyzed co-occur with reference to the storage means;
For each pair whose frequency calculated by the frequency calculation means exceeds the threshold, the frequency of the pair is included in the upper L (an integer greater than or equal to 2) rank among pairs having the same first word of the pair. A word specifying means for specifying the first word and the second word that make up the pair, except when there is not,
A graph display device comprising: a display unit configured to display a graph including the first word and the second word specified by the word specifying unit and a link connecting them.   The graph display device according to claim 1, wherein the subset deriving unit derives a subset in descending order of frequency of corresponding words in the subset.   The graph display device according to claim 1, wherein the storage unit stores one or a plurality of attributes for each sentence unit, and narrows down a sentence unit set to be analyzed according to the attribute.   The graph display device according to claim 1, wherein the sentence unit includes one or more sentences.   The graph display device according to claim 1, wherein the word is a noun.   The graph display device according to claim 1, wherein the word specifying unit prohibits the same word from being specified more than a predetermined number of times.   The graph display device according to claim 9, wherein the word specifying unit prohibits the same word from being specified a plurality of times. In a graph display program for displaying words that appear in a set of sentence units to be analyzed including a plurality of sentence units,
Storage means for storing words appearing in the sentence unit in association with the sentence unit;
Frequency calculation means for calculating the frequency in the sentence unit set with respect to words appearing in the sentence unit in the sentence unit set with reference to the storage means;
Referring to the frequency calculated by the frequency calculation means, a word specifying means for specifying the uppermost L (an integer of 2 or more) words in the sentence unit set;
For each word specified by the word specifying means, a subset derivation means for deriving a subset consisting of sentence units including the word from the sentence unit set,
As a display means for displaying a link from the word corresponding to the subset and the word corresponding to the derivation sentence unit set to the word corresponding to the derivation subset,
A program for graph display, characterized by causing a computer to function, and further causing the subset derivation means to function so as to derive a repeated subset starting from a root sentence unit set.