JP2015125594A - Information processing device, information processing method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technology enabling a user to easily find information intended by a user or variable information from a large-scale text.SOLUTION: An information processing device displays a list of words obtained by analyzing a document, receives selection of a word subjected to list display, creates a network graph containing words having received selection, and displays the graph. The network graph comprises a node where a word is displayed, and an edge connecting the nodes. When the node is selected, a word related to the word of the node is further displayed.

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program.

  On the WEB, enormous text data such as blog articles and SNS, papers and patent documents are generated one after another. It is required to quickly find valuable information or information intended by the user from this large-scale text as required.

  However, even if interesting or valuable information is generated, it will be overlooked unless SNS or papers are frequently confirmed.

  In addition, when searching for information from a large-scale text, a search word for accurately finding the intended information may not be easily determined. Furthermore, a simple search word may be caught in a search other than the information of the intended information, and it may be difficult to select information.

  In response to these problems, various methods for extracting central meanings and topics from large-scale texts have been proposed.

  The technique described in Patent Document 1 performs morphological analysis of text, creates a semantic network expressing words and concepts as nodes and edges as co-occurrence relationships between words and concepts, and creates words with strong co-occurrence relationships. This is a technique for finding more information by broadening the range of search words by finding out the concept.

  The technique described in Patent Document 2 uses a network graph that uses words and word groups obtained by morphological analysis of text as nodes, and in addition to information on co-occurrence relationships between nodes, the meanings such as synonyms, synonyms, and hierarchical relationships This is a technique that supports search and analysis by displaying a represented edge in a network graph and presenting a larger amount of information to improve readability compared to a network graph having only co-occurrence relationships.

JP 2001-243223 A JP 2009-75881 A

  The methods described in Patent Document 1 and Patent Document 2 are methods for extracting meanings and topics from large-scale texts, and contribute to solving the above-described problems.

  However, each method merely mechanically creates and presents a network graph, and there is no function that allows the user to select information.

  Further, although it is stated that the search range is expanded from the relationship between the nodes of the network graph, it is not discussed from the viewpoint of narrowing down the search range.

  Therefore, an object of the present invention is to provide a technique for making it easier for a user to find information intended by the user or valuable information from a large-scale text.

  The present invention provides a display control means for displaying a list of words obtained by analyzing a document, a word selection receiving means for receiving a selection of words displayed by the display control means, and a selection by the word selection receiving means. An information processing apparatus comprising: a creating unit that creates a network graph including the accepted word, wherein the display control unit displays the network graph created by the creating unit.

  Further, the present invention provides a display control step in which the display control means displays a list of words obtained by analyzing the document on the information processing apparatus, and a word selection accepting means of the information processing apparatus displays the list in the display control step. A word selection receiving step for receiving selection of the displayed word, and a creation step for creating a network graph including a word for which the creation means of the information processing apparatus has accepted the selection in the word selection receiving step, the display The control process displays the network graph created by the creation process.

  The present invention is also a program executable in an information processing apparatus, wherein the information processing apparatus displays a list of words obtained by analyzing a document, and the display control means displays a list. A word selection accepting unit that accepts selection of the selected word, and a creation unit that creates a network graph including the word accepted by the word selection accepting unit. The display control unit is a network created by the creating unit. A program characterized by functioning as a means for displaying a graph.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to provide the technique for making a user easy to find the information which a user intends or valuable information from a large-scale text.

It is a figure which shows the structure of the text narrowing-down system in embodiment of this invention. It is a figure which shows the hardware constitutions of the information processing apparatus in embodiment of this invention. It is an example of the text data in embodiment of this invention. It is an example of the analysis result data in the embodiment of the present invention. It is a flowchart which shows the analysis process in embodiment of this invention. It is a flowchart which shows the connection process of the morpheme in embodiment of this invention. It is an example of the morphological analysis and dependency analysis in embodiment of this invention. It is an example which shows the whole screen in which the user in embodiment of this invention browses and operates. It is an example which shows the modal window in embodiment of this invention. It is an example which shows the total result in the embodiment of the present invention. It is an example of the ranking table | surface in embodiment of this invention. It is an example which shows the ranking table which changed in embodiment of this invention. It is a flowchart which shows the process at the time of the word selection in embodiment of this invention. It is an example of the network graph in embodiment of this invention. It is an example of click operation in the embodiment of the present invention. It is an example of click operation in the embodiment of the present invention. It is an example of pinning operation in the embodiment of the present invention. It is an example of search operation in the embodiment of the present invention. It is an example of the narrowing-down result display in the embodiment of the present invention. It is an example of the calculation formula for calculating the raise degree of the word appearance frequency in embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a diagram showing a functional configuration of a text narrowing system according to an embodiment of the present invention. The text narrowing system 100 includes a database 110, an analysis processing unit 120, a totaling unit 130, a WEB application server 140, and a client terminal 150 operated by a user.

  The user can issue an instruction to the WEB application server 140 via the client terminal 150. Then, the WEB application server 140 transmits various data to the client terminal 150 based on the instruction received from the client terminal 150. The client terminal 150 displays the received various data on the display unit.

  The database 110, the analysis processing unit 120, the totaling unit 130, the WEB application server 140, and the client terminal 150 in the text narrowing system of the present invention have all the functions provided by one terminal even if each terminal has another function. Either of these functions may be provided. That is, the text narrowing system may be a single device (information processing device) as a whole, or may be a system divided into a plurality of terminals (information processing devices).

  The database 110 includes text data 111 for storing large-scale text (documents), and analysis result data 112 for storing frequency information of words or connected words that are morphological analysis and dependency analysis results of text and dependency information, and dependency information. . A connected word is a combination of words. Hereinafter, a word or a connected word is collectively referred to as a word.

  FIG. 2 is a diagram illustrating a hardware configuration of an information processing apparatus applicable to the client terminal 150 and the WEB application server 140.

  Hereinafter, an example of the hardware configuration of the information processing apparatus in the text narrowing system shown in FIG. 1 will be described with reference to FIG.

  In FIG. 2, reference numeral 201 denotes a CPU that comprehensively controls each device and controller connected to the system bus 204. Further, the ROM 203 or the external memory 211 is necessary to realize a BIOS (Basic Input / Output System) or an operating system program (hereinafter referred to as an OS), which is a control program of the CPU 201, or a function executed by each server or each PC. Various programs are stored.

  A RAM 202 functions as a main memory, work area, and the like for the CPU 201. The CPU 201 implements various operations by loading a program necessary for execution of processing from the ROM 203 or the external memory 211 into the RAM 202 and executing the loaded program.

  An input controller 205 controls input from the input device 209 or the like. A video controller 206 controls display on a display device 210 such as a liquid crystal display. The display device is not limited to a liquid crystal display, and may be a CRT display or the like. These are used by clients as needed.

  A memory controller 207 is connected to the hard disk (HD), flexible disk (FD), or PCMCIA card slot for storing a boot program, various applications, font data, user files, editing files, various data, etc. via an adapter. The access to the external memory 211 such as a compact flash (registered trademark) memory is controlled.

  A communication I / F controller 208 connects and communicates with an external device via a network, and executes communication control processing on the network. For example, communication using TCP / IP is possible.

  Note that the CPU 201 enables display on the display device 210 by executing outline font rasterization processing on a display information area in the RAM 202, for example. Further, the CPU 201 enables a user instruction with a mouse cursor (not shown) on the display device 210.

  Various programs that operate on the hardware are recorded in the external memory 211 and are executed by the CPU 201 by being loaded into the RAM 202 as necessary.

  It should be noted that not all apparatuses have these configurations, and it is only necessary to have necessary ones.

  FIG. 3 is a diagram illustrating an example of the text data 111. The text data 111 holds, as tags, an id for uniquely identifying the data, a created date and time, a text title, a body, a URL, a medium on which the text data is collected, a search key used for the collection, and the like. Tags can be freely set even when there are multiple or no tags when collecting text data. For example, when text data relating to a camera is collected from an SNS on the network, a tag is set such as “SNS, camera”. By setting tags, the data used for analysis is narrowed down.

  When collecting text data, if there is a dedicated data collection API, the API is executed periodically (batch processing, differential processing) to collect data, search keywords, media names (SNS names, etc.), etc. Is registered as a tag.

  For example, data collected by the search word “Canon, camera” using the Twitter API is attached with a tag (list) of [Canon, camera, twitter], and the title is blank and stored in the DB.

  In addition, for data on the WEB that does not have a data collection API, RSS or HTML / XML analysis of the corresponding page is performed and data is collected regularly, and the media name (site name, blog name, etc.) is used as a tag. Store in DB.

  For example, data collected from RSS of a news summary site is tagged with a “news article” and the title and URL are stored together in the DB.

  FIG. 4 is a diagram illustrating an example of the analysis result data 112. As an example, an analysis result example of data of id = 1 in FIG. 3 is shown. The analysis result data 112 is an id indicating which data of the text data 111 is analyzed, a dependency relationship and dependency relationship frequency extracted by the analysis processing unit, noun information and noun information frequency, adjective information and adjective. Information frequency, verb information, and verb information frequency are held. The adjective information and verb information hold the final form of the morpheme together with the original text acquired from the text. The reason why the adjectives and verbs hold the final form is to aggregate the variation of expressions in the use of morphemes and to summarize the frequency of the same morphemes. Details of the analysis processing unit will be described later.

  Next, analysis processing in the analysis processing unit 120 will be described with reference to FIG. According to the process shown in FIG. 5, the analysis processing unit 120 acquires the text body from the text data 111 and executes the analysis, aggregates the analysis results, and registers or updates the analysis result data 112. It is assumed that at the time of text data acquisition, only differential data that has not been analyzed is acquired, and the analysis processing is automatically executed periodically at intervals of 30 minutes.

  In step S501, the text body is acquired from the text data 111 and divided into sentences. Sentences between sentences are judged by using punctuation marks, full-width spaces, closing parentheses, etc.

  Then, the processes in steps S503 to S506 are repeatedly executed for each sentence divided in step S501 (step S502).

  In step S503, morphological analysis and dependency analysis are executed, and the sentence is divided into clauses and morphemes. Then, the morpheme part-of-speech, end form, and dependency information between phrases are acquired.

  Here, FIG. 7 shows an analysis example of a sentence “I bought this product because it was expensive but very good”. Note that morphological analysis and dependency analysis can be implemented using existing technology, and thus detailed description thereof is omitted here.

  As shown in Fig. 7, the sentence "I bought this product because it was expensive but very good" was "It was expensive", "Very", "It's good", "This", "Product" , It is divided into six clauses "I bought it."

  In addition, for example, the phrase “It was high” is divided into three morphemes, “High”, “Ta”, and “Bad”, and each part of speech is an adjective, auxiliary verb, and connective particle. ”,“ Ta ”,“ Bad ”.

  In addition, it shows that the clause of the dependency of “It was expensive” is “Because it is good”.

In addition, although the clause of the dependency destination of “I bought” is “None”, this indicates that there is no clause of the dependency destination.

  Although the results of similar analysis are shown for other clauses, detailed description is omitted.

  Next, the processes in steps S505 to S506 are repeatedly executed for each phrase acquired in step S503 (step S504).

  In step S505, a morpheme concatenation process is executed for each phrase.

  Here, the morpheme concatenation process in step S505 will be described with reference to the flowchart of FIG.

  First, the morphemes in the phrase are sequentially scanned, and the processes in steps S602 to S604 are executed for each morpheme.

  In step S602, the morpheme part of speech is identified. If it is an adjective or verb, the adjective / verb concatenation process in step S603 is performed. If it is a noun, the noun concatenation process in step S604 is performed. Move to next morpheme processing.

  In step S603, if the morpheme part of speech following the morpheme to be processed is an adjective or verb or a particle excluding a case particle, morpheme concatenation is repeatedly executed. Finish the process.

In step S604, if the part of speech of the morpheme that follows the morpheme to be processed is a noun, the morpheme is repeatedly connected. If the end of the phrase or the connection condition is not satisfied, the process of this flowchart ends.
Returning to the description of FIG.

In step S506, the end forms of morphemes obtained in steps S103 and 105, concatenated words, and dependency frequencies are totaled, a list is created, and the analysis result data 112 is registered / updated.
Then, the process of the flowchart of FIG.

  The totaling unit 130 acquires the analysis result data 112 desired by the user from the database 110 using the date and time and tag specified by the user according to the instruction to the WEB application server 140, and returns the totaling result to the WEB application server 140.

  Next, a screen (screen displayed on the client terminal 150) presented to the user of the text narrowing system 100 by the WEB application server 140 and an operation method by the user will be described.

  FIG. 8 shows an example of an entire screen where the user searches for a search word for narrowing down information and browses the narrowed down information.

The entire screen includes an analysis setting button 200 for setting a tag and a date for selecting an analysis result to be read, a dependency, a noun, an adjective and a verb ranking table 300 (an area in which words are displayed in a list), A network graph 400 that can be operated in real time and a narrowing result display column 500 that presents information narrowed down by a search word selected from the network graph (original text of a document including the word) are included.
The analysis setting button 200 will be described with reference to FIG.

  When the analysis setting button 200 is pressed by the user, a modal window 900 shown in FIG. 9 is displayed on the screen shown in FIG.

When the user selects a date and time and a tag in the modal window 900 and presses a read button, the totaling unit 130 analyzes the result desired by the user from the database 110 (result specified by the selected date and tag). Get the word rise and. Then, a total result obtained by totaling the acquired analysis results is returned to the WEB application server 140.
The explanation of the degree of increase and the total result will be described later.

  Then, the WEB application server 140 creates the contents to be presented in the ranking table 300 and the graph structure data to be drawn on the network graph 400 from the information acquired from the totaling unit 130.

  The contents presented in the ranking table 300 and the creation of the network graph 400 will be described later.

  Note that the modal window 900 is closed when the read button is pressed or a place other than the modal window 900 is selected by left-clicking the mouse or the like.

  Next, an example of the degree of increase and the total result will be described with reference to FIG. The tabulation result is information such as dependency relationship, noun information, adjective information, verb information, each word, its frequency, and degree of increase, which is obtained by tabulating information acquired from the analysis result data 112 by the tabulation unit 130.

  The degree of increase is calculated by calculating the ratio of the appearance frequency (number of appearances) of the word between the period specified by the user in the modal window 900 and the comparison period, so that how many words are in the specified period compared to the comparison period. It is an index that indicates whether it has appeared. The comparison period is the entire period, the previous month, the previous week, the previous day, or the like.

  An example of the formula for calculating the degree of increase is shown in FIG. 20, but is not limited to this, and can be calculated by other calculation methods.

The ranking table 300 is a list in which words are presented in the order of frequency or increasing degree for each of the dependency, the noun, the adjective, and the verb using the counting result acquired from the counting unit 130.
The ranking table 300 will be described with reference to FIG.

  Although FIG. 11 shows an example of the adjective ranking table, the same is true for the dependency, noun, and verb ranking tables.

  First, the display contents of the ranking table 300 will be described. In the ranking table 300, a word column for displaying words (in the example of FIG. 11, “good”, “want”, “high”, etc. are displayed) and a frequency column for displaying the appearance frequency of words (in the example of FIG. 11). , “Good” has a value of 317, “Want” has a value of 256, etc.) and a degree column for displaying the degree of word rise (in the example of FIG. 11, “Good” is a horizontal arrow (horizontal Arrow), “desired” is indicated by an upward arrow, and “high” is indicated by a downward arrow).

  In the word column of the adjective and verb ranking table 300, the word end form is displayed in advance, and if there is a connected word, information on the connected word is displayed as necessary. The display of the connected words in the adjective and verb ranking table 300 will be described later.

  In the word column of the dependency and noun ranking table 300, information on connected words is displayed from the beginning. By displaying an image such as an arrow according to the value of the degree of increase in the degree of increase column, the ratio with the comparison period is presented in an easy-to-understand manner. For example, when the degree of increase is 0 to 1, and the closer to 1, the more frequently the words or connected words in the specified period are, the upward degree is less than 0.4 and the upward degree is 0.4 or more. When the value is less than 0.7, a horizontal arrow is displayed, and when the degree of increase is 0.7 or more, an upward arrow is displayed.

  Next, functions of the ranking table 300 will be described. The ranking table 300 displays several words information having a higher frequency. Word information that cannot be displayed is a pagination button 320 that can be browsed sequentially by pagination, a number display 310 that displays information on what is displayed in the ranking table 300 in the whole, and frequency of all words Alternatively, a sort button 330 for sorting the display contents of the ranking table 300 by sorting in ascending or descending order of the degree of increase and a search window 340 for searching for words in the ranking table 300 are provided.

  Further, when the word column of the ranking table 300 is selected by left-clicking the mouse or the like, the selected word is added as a node of the network graph 400 shown in FIG.

  However, in the case where the ranking table 300 is an adjective or a verb, when a word is selected, information of a connected word associated with the selected word is displayed as shown in FIG. It is added as 400 nodes.

  FIG. 12 shows an example when “good” is selected in the word column of FIG. It is assumed that the original display is restored when the table in FIG. 12 is selected by right-clicking the mouse or the like.

  Details of the node addition processing to the network graph 400 by the network graph 400 and word selection will be described later.

  Next, the network graph 400 will be described. The network graph 400 is a graph in which a word is a node and an edge is added between nodes of a word having a dependency relationship based on the aggregation result acquired from the aggregation unit 130.

  The size and color of the node and the color and thickness of the edge are determined based on the information of the word and the dependency relationship.

  For example, the diameter of the node circle is determined based on the appearance frequency of the word, such as increasing the diameter of the node circle when the appearance frequency of the word is high. Also, the node color is determined according to the type of part of speech. Further, the thickness of the edge is determined based on the frequency of the dependency relationship between the words connected by the edge (for example, the thickness is increased if the frequency is high). In addition, in consideration of the order of the dependency relationship, the graph may be displayed as an effective graph with the edges directed.

  Next, processing when a word in the ranking table 300 is selected will be described with reference to FIG.

  In step S1301, information on the word selected from the counting result is acquired.

  In step S1302, it is determined whether a dependency relationship exists based on the information acquired in step S1301.

  If there is a dependency relationship (step S1302: YES), the process proceeds to step S1303.

  If there is no dependency relationship (step S1302: NO), the process proceeds to step S1304.

  In step S1303, the process of step S1303 is executed in a predetermined number of dependency relationships, and the nodes and edges are displayed on the network graph 400 based on the information of the aggregation results.

If there is no dependency relationship (S1304), the selected word is displayed as a node.
The predetermined number of times is a threshold value for limiting the number of edges to be displayed in advance in order of decreasing dependency frequency so that the number of nodes and edges to be displayed does not become excessive. FIG. 14 is a display example on the network graph 400 when “good” is selected from the adjective ranking table of FIG.

  As shown in FIG. 14, three edges extend from a node “good because” and are connected to nodes “very”, “bought”, and “high”, respectively.

  Next, reception of user operations in the network graph 400 and functions corresponding thereto will be described.

There are roughly three functions: a node / edge addition / deletion function, a node pinning function, and a node search word addition / deletion function. Each function will be described below with reference to the drawings. First, a function for adding / deleting nodes and edges will be described.

  When the user accepts an operation such as selecting a node by, for example, left-clicking the mouse, the word of the selected node is acquired, and the word selection process shown in the flowchart of FIG. 13 is executed.

  With the process shown in FIG. 13, nodes and edges are additionally displayed a predetermined number of times, and nodes and edges that are more than a predetermined distance away from the selected node are automatically deleted from the network graph 400.

  The predetermined distance is a threshold when one edge is set as the distance 1 so that the number of edges and nodes displayed on the network graph 400 does not become excessive. Further, the positions of nodes and edges are automatically corrected according to the display frame of the network graph 400.

  Hereinafter, this operation (a predetermined operation for adding a node and an edge, such as an operation in which the user selects a node with a left click of the mouse) will be referred to as a click operation.

  An example of the click operation when the predetermined number of times and the predetermined distance are 3 will be described.

  FIG. 15 shows an example when the “very” node of FIG. 14 is clicked.

  Nodes of “Reputation” and “Beautiful” that have a dependency relationship with “Very” and additional edges are displayed, and the entire position is automatically corrected.

  Furthermore, FIG. 16 shows an example when the node “beautiful” in FIG. 15 is clicked.

The nodes “photograph” and “capable of taking” that are related to “beautiful” and the edges connected to them are additionally displayed, the nodes “bought” and “high” that are more than the predetermined distance and the edges connected to it Is automatically deleted, and the entire position is automatically corrected. That is, when the number of edges connecting from a terminal node to another terminal node exceeds a predetermined number, a network graph is created by deleting the terminal nodes so that the predetermined number is reached.
Next, the node pinning function will be described.

  The pinning function is a function that can be pinned by receiving an operation in which a user selects a node by, for example, right-clicking the mouse.

  When pinning is performed, edges and nodes existing between a pinned node and a newly selected node are not automatically deleted even if they are separated by a predetermined distance or more.

  That is, even when the number of edges connecting from the terminal node to the other terminal node exceeds a predetermined number, an instruction (pinning operation) is given to continue displaying the terminal node. In this case, the network graph is created without deleting the node.

  A pinned icon is displayed at the top of the node that has been pinned, and pinning can be canceled by selecting it again.

  Hereinafter, this operation (a predetermined operation for pinning a node, such as an operation in which the user selects a node by right-clicking the mouse) will be referred to as a pinning operation.

  An example of the pinning operation when the predetermined number of times and the predetermined distance are 3 will be described.

  FIG. 17 shows an example in which the “clean” node is clicked after the “buy” node in FIG. 15 is pinned, and a pinning icon 410 is displayed above the “buy” node. Is displayed. Further, unlike the network graph shown in FIG. 16, the edges and nodes existing between “beautiful” and “bought” remain displayed without being automatically deleted.

  Next, the node search word addition / deletion function will be described.

  When the user accepts an operation such as selecting a node with a left double click of the mouse, for example, the word of the selected node is registered as a search word.

  Then, the text data 111 is searched using the registered search word, and the search result is displayed in the narrowed-down result display column 500.

  A plurality of words can be registered in the search word, and a search word registration icon (420 in FIG. 18) is displayed below the node registered in the search word, and the registration can be deleted by selecting again. Can do. When the selected node is deleted from the network graph 400 due to restrictions such as a predetermined distance, the corresponding word is automatically deleted from the registered search word. Hereinafter, this operation (a predetermined operation for registering a search word such as a user selecting a node with a left double click of the mouse) will be referred to as a search operation.

  An example of a search operation will be described. FIG. 18 shows an example of a network graph when searching for the nodes “beautiful”, “captured” and “purchased” in FIG. 17, and the nodes of the nodes “clean”, “capable” and “buy” are shown. A search word registration icon 420 is displayed at the bottom. Then, using the registered search word and the period and tag specified by the user in the modal window 900, the text “111”, “Take” and “Buy” words from the text data 111 are present in the text body or title. Data is acquired and displayed in a list in the refined result display field 500 (FIG. 19).

  Next, the narrowing-down result display column 500 will be described with reference to FIG.

  The refinement result display field 500 includes a title field for displaying the title of the text data 111, a text field for displaying the text body, and a tag field for displaying a tag. In the title column, if there is a word that matches the search word, the word is highlighted. If a URL exists, the title is displayed with a URL link.

  As an example of highlighting, for example, a matching word is highlighted and displayed. In the text column, the text to be displayed is restricted so as not to be too long, and is highlighted so that a word that matches the search word can be seen.

  For example, only the sentence including the search word is extracted from the text body, and the word that matches the search word is highlighted and displayed.

  In addition, a designated period display 510 for displaying a period designated by the user in the modal window 900, a number display 520 for displaying information on what items are displayed in the narrowed-down result display column 500 among all search results, The pagination button 530 for sequentially browsing the search results that could not be displayed by pagination, the sort button 540 for sorting by the tag specified by the user in the modal window 900, the title of the text, and the refined result display field 500 A search window 550 is provided for searching for.

  By using this narrowing-down result display column 500, the user can confirm intended information or interesting information from a large-scale text, and can acquire information from the large-scale text at a reduced cost for searching for information. Can do.

  As described above, according to the present invention, the user of the text narrowing system 100 can use the system by making settings through the modal window 900, and the ranking table 300, the network graph 400, and the narrowing result display. By mutually confirming and operating the columns 500, it is possible to find a search word for acquiring information desired by itself, and to efficiently find relevant information from a large-scale text.

  In the above description, the node coloring example in the network graph 400 is the type of part of speech, but the information is affirmed by coloring and presenting the evaluation polarities such as negative elements and positive elements of the word. It can be used to judge whether the content is specific or negative.

  In addition, in the text data 111, by retaining the user attributes such as the age and sex of the person who created the text data, the user attributes are displayed in the refined result display column 500, and information such as age and gender is displayed. Can be presented together.

  It should be noted that the configuration and contents of the various data described above are not limited to this, and it goes without saying that the various data and configurations are configured according to the application and purpose.

  The program in the present invention is a program for causing a computer to execute the processes of FIGS. Note that the program in the present invention may be a program for each process in FIGS. 5, 6, and 13.

  As described above, a recording medium that records a program that implements the functions of the above-described embodiments is supplied to a system or apparatus, and a computer (or CPU or MPU) of the system or apparatus stores the program stored in the recording medium. It goes without saying that the object of the present invention can also be achieved by reading and executing.

  In this case, the program itself read from the recording medium realizes the novel function of the present invention, and the recording medium recording the program constitutes the present invention.

  As a recording medium for supplying the program, for example, a flexible disk, hard disk, optical disk, magneto-optical disk, CD-ROM, CD-R, DVD-ROM, magnetic tape, nonvolatile memory card, ROM, EEPROM, silicon A disk or the like can be used.

  Further, by executing the program read by the computer, not only the functions of the above-described embodiments are realized, but also an OS (operating system) operating on the computer based on an instruction of the program is actually It goes without saying that a case where the function of the above-described embodiment is realized by performing part or all of the processing and the processing is included.

  Furthermore, after the program read from the recording medium is written to the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, the function expansion board is based on the instructions of the program code. It goes without saying that the case where the CPU or the like provided in the function expansion unit performs part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.

  The present invention may be applied to a system constituted by a plurality of devices or an apparatus constituted by a single device. Needless to say, the present invention can be applied to a case where the present invention is achieved by supplying a program to a system or apparatus. In this case, by reading a recording medium storing a program for achieving the present invention into the system or apparatus, the system or apparatus can enjoy the effects of the present invention.

  Furthermore, by downloading and reading a program for achieving the present invention from a server, database, etc. on a network using a communication program, the system or apparatus can enjoy the effects of the present invention. In addition, all the structures which combined each embodiment mentioned above and its modification are also included in this invention.

100 Text Refinement System 110 Database 111 Text Data 112 Analysis Result Data 120 Analysis Processing Unit 130 Totaling Unit 140 WEB Application Server 150 Client Terminal

Claims (12)

Display control means for displaying a list of words obtained by analyzing the document;
Word selection accepting means for accepting selection of words displayed in a list by the display control means;
Creating means for creating a network graph including a word accepted by the word selection accepting means;
With
The information processing apparatus, wherein the display control means displays the network graph created by the creating means.   2. The network graph created by the creation means is a diagram composed of a node displaying the word and an edge indicating a dependency relationship between the word and the word and connecting the nodes. The information processing apparatus described in 1. The word selection accepting means further accepts selection of a word displayed on the network graph created by the creating means,
When the creation unit receives a selection of a word displayed on the network graph by the word selection reception unit, a network in which a node of a word having a dependency relationship with the word and an edge connecting the node are added The information processing apparatus according to claim 2, wherein a graph is created.   4. The information according to claim 3, wherein the display control means further displays the original text of the document including the word when the selection of the word displayed on the network graph is accepted by the word selection accepting means. Processing equipment.   5. The information processing apparatus according to claim 2, wherein a node in the network graph has a different size according to an appearance frequency of a word displayed in the node.   The information processing apparatus according to claim 2, wherein the edge of the network graph has a different thickness depending on a frequency of dependency between words connected by the edge.   The information processing apparatus according to claim 4, wherein the display control unit displays a list of words, a network graph, and an original text of a document on the same screen.   3. The creation unit further creates a network graph capable of identifying the word that has received the selection when the selection of the word displayed on the network graph is received by the word selection reception unit. The information processing apparatus according to any one of 1 to 7.   The creation unit further creates a network graph in which the end nodes are deleted so that the number of edges connecting from the end node to another end node exceeds a predetermined number so as to be within the predetermined number. The information processing apparatus according to claim 2, wherein the information processing apparatus is an information processing apparatus.   The creation means is further instructed to continue displaying the end node even when the number of edges connecting from the end node to another end node exceeds a predetermined number. The information processing apparatus according to claim 8, wherein the node is not deleted. A display control step in which the display control means in the information processing apparatus displays a list of words obtained by analyzing the document; and
A word selection receiving unit of the information processing apparatus for receiving a selection of words displayed in a list by the display control step;
A creation step of creating a network graph including a word whose selection is accepted by the word selection acceptance step by the creation means of the information processing device;
With
The display control step displays the network graph created by the creation step. A program executable in the information processing apparatus,
The information processing apparatus;
Display control means for displaying a list of words obtained by analyzing the document;
Word selection accepting means for accepting selection of words displayed in a list by the display control means;
Function as a creation means for creating a network graph including a word accepted by the word selection acceptance means,
A program for causing the display control means to function as means for displaying a network graph created by the creating means.

Images