TWI703457B - Text exploration method, text exploration program and text exploration device - Google Patents

Abstract

The text mining method of the present invention has the following steps: a step of extracting words from text data; a step of generating a symbiosis matrix from the extracted; a step of generating a symbiosis network based on the generated symbiosis matrix; and displaying the generated symbiosis The steps of the network screen. When inputting instructions for specifying the word of interest in the first screen of the first symbiosis network containing the entirety of the specified text data, extract words from the limited text data that includes the part of the specified text data that contains the word of interest , Use limited text data for the extracted words to generate a second co-occurrence matrix, generate a second co-occurrence network based on the second co-occurrence matrix, and display the second screen containing the second co-occurrence network.

Description

Text exploration method, text exploration program and text exploration device

The present invention relates to text mining, in particular to text mining methods, text mining programs, and text mining devices that display pictures of symbiotic networks containing words.

In recent years, the analysis of freely written text data and the search for useful information based on the analysis results have attracted attention. In text exploration, for example, by extracting words from the text data of the analysis object, and analyzing the appearance frequency or tendency of the words, the information is obtained.

When analyzing freely described textual data, the analyst does not subjectively select objects in the initial stage, but must grasp the overall situation of the textual data. Therefore, there are cases where an analyst uses a symbiotic network of words contained in text data.

Fig. 19 is a diagram showing an example of a symbiotic network. The symbiosis network extracts more pairs of words contained in the same sentence from text data, and uses undirected graphs to express the results. When the word Wa and word Wb in the text data of the analysis object are mostly contained in the same sentence, the symbiosis network includes the node corresponding to the word Wa, the node corresponding to the word Wb, and connecting the two The side. The symbiosis network shown in Fig. 19 includes a node corresponding to "staff", a node corresponding to "corresponding", and an edge connecting the two. Observing the symbiosis network shown in Fig. 19, it can be seen that in the text data of the analysis object, "staff" and "corresponding" are mostly contained in the same sentence.

Generally speaking, the symbiosis network is generated based on the whole specified text data. Hereinafter, this kind of symbiosis network is referred to as the "holistic symbiosis network." The analyst selects a plurality of words that should be paid attention to from the overall symbiosis network according to the hypothesis or analysis purpose established by himself (hereinafter referred to as the word of interest), and considers the words of interest for subsequent analysis.

When the analyst selects the word of interest, in order to determine whether the selected word of interest is suitable for the purpose of analysis, etc., he examines how to use the word of interest in the sentence containing the word of interest. Therefore, analysts sometimes use a symbiosis network based on text data (hereinafter referred to as limited text data) including sentences containing words of interest in the specified text data. Furthermore, the "article containing the word of interest" mentioned here does not only refer to the case of a single sentence containing the word of interest. It means that the paragraph including the sentence containing the word of interest is divided into plural sentences (sentences). Of the collection). Hereinafter, this kind of symbiosis network is referred to as a "limited symbiosis network." The analyst can grasp the content of the limited text data by using the limited symbiosis network. The analyst repeatedly refers to the overall symbiosis network and the limited symbiosis network until all the words of interest are selected.

In the following, consider generating a symbiosis network of words contained in text data, and a text exploration device that displays a screen containing the generated symbiosis network. In Japanese Patent Laid-Open No. 8-314980, there is a document database display device that generates an overall symbiosis network for each of a plurality of documents, and displays a screen containing the generated plurality of overall symbiosis networks. The display device retrieves the words input by the user from a plurality of overall symbiosis networks, and highlights the retrieved words on the screen.

The conventional text exploration device generates a symbiosis network based on the whole specified text data. Therefore, according to the conventional text exploration device, the screen including the overall symbiotic network can be easily displayed.

On the other hand, when using a conventional text mining device to display a screen containing a limited symbiotic network, the analyst must perform complicated operations. Specifically, every time an analyst selects a word of interest from the overall symbiosis network, he must generate limited text data based on the specified text data, and assign the generated limited text data to the text exploration device. In addition, the analyst refers to both the overall symbiosis network and the limited symbiosis network when selecting the words of interest. Therefore, the text mining device must store both the image data of the overall symbiosis network and the image data of the limited symbiosis network. However, when more symbiotic networks are generated, the preservation and management of image data becomes difficult.

Therefore, the object of the present invention is to provide a text mining method, text mining program, and text mining device that can display the screen of the symbiosis network containing the specified word of interest with simple operations.

The first aspect of the present invention is a text exploration method, which displays a screen containing the analysis results of text data; it is characterized in that it has the following steps: a step of extracting words from text data; and generating a co-occurrence matrix for the words The steps of; the steps of generating a symbiosis network based on the above symbiosis matrix; and the steps of displaying the screen containing the above symbiosis network; enter the designation in the first screen of the first symbiosis network including the entirety based on the specified text data When the word of interest is instructed, the step of extracting the word is to extract the word from the limited text data that includes the part of the word of interest in the specified text data, and the step of generating the co-occurrence matrix is to apply the restriction to the word The text data generates a second symbiosis matrix, the step of generating the symbiosis network is based on the second symbiosis matrix to generate a second symbiosis network, and the step of displaying the screen is to display a second screen containing the second symbiosis network.

The second aspect of the present invention is in the first aspect of the present invention, and is characterized in that by selecting one or more nodes included in the first symbiosis network in the first screen, and selecting the analysis start , And the input indicates that the word corresponding to the above node is designated as the above-mentioned word of interest.

The third aspect of the present invention is in the first aspect of the present invention, and is characterized in that by continuing to select a node included in the first symbiosis network in the first screen, the input will be the same as the node The corresponding word is designated as an indication of the above-mentioned word of interest.

The fourth aspect of the present invention is in the first aspect of the present invention, and is characterized in that by continuing to select an edge included in the first symbiosis network in the first screen, the input will be connected to the The words corresponding to the two nodes of the edge are designated as the indication of the above-mentioned word of interest.

The fifth aspect of the present invention is in the first aspect of the present invention, characterized in that by selecting one or more edges included in the first symbiosis network in the first screen, and selecting the start of analysis, The input indicates that the words corresponding to the plural nodes connected to the aforementioned edges are designated as the aforementioned words of interest.

The sixth aspect of the present invention is in the first aspect of the present invention, and is characterized in that when the merge instruction is entered in the second screen including a plurality of second symbiotic networks, the step of displaying the above screen is to change the above The second symbiosis network is displayed as a label.

The seventh aspect of the present invention is in the sixth aspect of the present invention, and is characterized in that by clicking on a second symbiosis network in the second screen and releasing it in another second symbiosis network, Enter the above merge instructions.

The eighth aspect of the present invention is in the first aspect of the present invention, and is characterized in that the above-mentioned limited text data includes a sentence containing the aforementioned word of interest in the above-mentioned designated text data.

The ninth aspect of the present invention is in the eighth aspect of the present invention, and is characterized in that the above-mentioned limited text data when specifying a plurality of words of interest include all sentences in the specified text data that contain the plurality of words of interest. .

The tenth aspect of the present invention is in the eighth aspect of the present invention, and is characterized in that the above-mentioned limited text data when specifying a plurality of words of interest includes any one of the above-mentioned plurality of words of interest in the specified text data The sentence.

The eleventh aspect of the present invention is in the first aspect of the present invention, and is characterized in that the step of generating the co-occurrence matrix is to generate a co-occurrence matrix with Jaccard coefficients as elements.

The twelfth aspect of the present invention is a text exploration program, which is used to display a screen containing the analysis result of text data; it is characterized in that the CPU uses the memory to make the computer perform the following steps: extract words from text data The step of; the step of generating a symbiosis matrix for the aforementioned words; the step of generating a symbiosis network based on the aforementioned symbiosis matrix; and the step of displaying a screen containing the aforementioned symbiosis network; in the first symbiosis including the entirety based on the specified text data When inputting instructions for specifying the word of interest in the first screen of the network, the step of extracting the word is to extract the word from the limited text data that includes the part of the word of interest in the specified text data to generate the co-occurrence matrix The step is to generate a second symbiosis matrix using the limited text data for the above words, the step of generating the symbiosis network is to generate a second symbiosis network based on the second symbiosis matrix, and the step of displaying the screen is to display that the second symbiosis is included The second screen of the Internet.

The thirteenth aspect of the present invention is in the twelfth aspect of the present invention, and is characterized in that one or more nodes included in the first symbiosis network are selected in the first screen, and the analysis starts , And the input indicates that the word corresponding to the above node is designated as the above-mentioned word of interest.

The fourteenth aspect of the present invention is in the twelfth aspect of the present invention, and is characterized in that by continuing to select a node included in the first symbiotic network in the first screen, the input will be the same as the node The corresponding word is designated as an indication of the above-mentioned word of interest.

The fifteenth aspect of the present invention is in the twelfth aspect of the present invention, and is characterized in that by continuing to select an edge included in the first symbiosis network in the first screen, the input will be connected to the The words corresponding to the two nodes of the edge are designated as the indication of the above-mentioned word of interest.

The sixteenth aspect of the present invention is in the twelfth aspect of the present invention, and is characterized in that by selecting one or more edges included in the first symbiosis network in the first screen, and selecting the start of analysis, The input indicates that the words corresponding to the plural nodes connected to the aforementioned edges are designated as the aforementioned words of interest.

The seventeenth aspect of the present invention is in the twelfth aspect of the present invention, and is characterized in that when a merge instruction is entered in a second screen including a plurality of second symbiotic networks, the step of displaying the screen is to change the plurality of The second symbiosis network is displayed as a label.

The eighteenth aspect of the present invention is in the seventeenth aspect of the present invention, and is characterized in that by clicking on a second symbiosis network in the second screen and releasing it in another second symbiosis network, Enter the above merge instructions.

The 19th aspect of the present invention is a text exploration device that displays a screen containing the analysis result of text data; it is characterized in that it includes: a word extraction unit that extracts words from text data; and a symbiosis matrix generation unit , Which generates a symbiosis matrix for the above words; a symbiosis network generation section, which generates a symbiosis network based on the symbiosis matrix; and a screen display section, which displays a screen containing the symbiosis network; When inputting instructions to specify the word of interest in the first screen of the overall first symbiosis network, the word extraction unit extracts the word from the limited text data that includes the portion of the word of interest in the specified text data, and the symbiosis The matrix generating unit generates a second co-occurrence matrix using the limited text data for the word, the co-occurrence network generating unit generates a second co-occurrence network based on the second co-occurrence matrix, and the screen display unit displays the second co-occurrence network including the second co-occurrence network 2 pictures.

The twentieth aspect of the present invention is in the nineteenth aspect of the present invention, and is characterized in that when a merge instruction is input in a second screen including a plurality of second symbiotic networks, the screen display unit displays the plurality of second symbiosis networks. 2 The symbiosis network is displayed as a label.

According to the first, twelfth, or nineteenth aspect above, when inputting the instruction of the designated word of interest in the first screen of the first symbiosis network containing the entirety of the designated text data, the display contains the text based on the designated The second screen of the second symbiosis network of the part of the data that contains the word of interest. Therefore, it is possible to display the screen of the symbiosis network including the specified word of interest with a simple operation.

According to the second or 13th aspect above, by selecting one or more nodes in the first screen and starting the analysis, you can input instructions to specify one or more words of interest by simple operations, and display including the specified 1 The screen of the symbiosis network when one or more words follow.

According to the third or fourteenth aspect above, by continuing to select a node in the first screen, you can input instructions to specify a word of interest with a simple operation, and display the symbiosis network including the specified word of interest The picture.

According to the 4th or 15th aspect above, by continuing to select an edge in the first screen, you can input instructions for specifying 2 words of interest with a simple operation, and display the symbiosis network that contains the specified 2 words of interest. Picture.

According to the 5th or 16th aspect above, by selecting one or more edges in the first screen and starting the analysis, you can input instructions for specifying multiple words of interest with a simple operation, and display when the specified multiple words of interest are included The picture of the symbiosis network.

According to the sixth, 17th, or 20th aspect, by displaying a plurality of second symbiosis networks in label form when inputting a merge instruction, a plurality of second symbiosis networks can be displayed concisely.

According to the seventh or eighteenth aspect, by clicking or releasing the second symbiosis network in the second screen, a simple operation can be used to input the merge instruction, and a plurality of second symbiosis networks can be displayed concisely.

According to the eighth aspect above, when inputting instructions for specifying the word of interest, the specified text data can be divided into sentence units to obtain the limited text data, and the second symbiosis network containing the obtained limited text data can be displayed. Picture of the road.

According to the ninth or tenth aspect above, the screen including the second symbiosis network when AND processing or OR processing plural words of interest can be displayed.

According to the eleventh aspect, by generating a co-occurrence matrix with Jaccard coefficients as elements, the co-occurrence of words contained in the text data can be better analyzed.

10‧‧‧Text exploration device

11‧‧‧Instruction input section

12‧‧‧Text data storage department

13‧‧‧Word Extraction Section

14‧‧‧Symbiosis Matrix Generation Department

15‧‧‧Symbiosis Network Generation Department

16‧‧‧Screen display

20‧‧‧Computer

21‧‧‧CPU

22‧‧‧Main memory

23‧‧‧Storage Department

24‧‧‧Input part

25‧‧‧Display

26‧‧‧Ministry of Communications

27‧‧‧Recording media reading section

28‧‧‧Keyboard

29‧‧‧Mouse

30‧‧‧Recording media

31‧‧‧Text exploration program

32‧‧‧Text data

41~45‧‧‧Window

51‧‧‧Integral Symbiosis Network

52~54‧‧‧Limited symbiosis network

61‧‧‧Analyze button

62‧‧‧Mouse cursor

63~64‧‧‧label

71~75‧‧‧Screen

Fig. 1 is a block diagram showing the structure of a character exploration device according to an embodiment of the present invention.

Fig. 2 is a block diagram showing the structure of a computer functioning as the text mining device shown in Fig. 1.

Fig. 3 is a flowchart showing the operation of the text exploration device shown in Fig. 1.

Fig. 4 is a diagram showing an example of a co-occurrence matrix generated by the character exploration device shown in Fig. 1.

FIG. 5 is a diagram showing an example of a window including the overall symbiosis network displayed by the text mining device shown in FIG. 1.

FIG. 6 is a diagram showing the first operation of specifying the word of interest in the window shown in FIG. 5.

FIG. 7 is a diagram showing the second operation of specifying the word of interest in the window shown in FIG. 5.

FIG. 8 is a diagram showing the third operation of specifying the word of interest in the window shown in FIG. 5.

FIG. 9 is a diagram showing the fourth operation of specifying the word of interest in the window shown in FIG. 5.

FIG. 10 is a diagram showing the fifth operation of specifying the word of interest in the window shown in FIG. 5.

FIG. 11 is a diagram showing the sixth operation of specifying the word of interest in the window shown in FIG. 5.

FIG. 12 is a diagram showing an example of a window including a limited symbiotic network displayed by the text mining device shown in FIG. 1.

FIG. 13 is a diagram showing an example of a window including a limited symbiotic network displayed by the text mining device shown in FIG. 1.

Fig. 14 is a diagram showing an example of a display screen of the character exploration device shown in Fig. 1.

Fig. 15 is a diagram showing an example of a display screen of the character exploration device shown in Fig. 1.

Fig. 16 is a diagram showing an example of a display screen of the character exploration device shown in Fig. 1.

FIG. 17 is a diagram showing the operation of merging windows in the text exploration device shown in FIG. 1.

FIG. 18 is a diagram showing the display screen after the operation shown in FIG. 17 is performed.

Fig. 19 is a diagram showing an example of a symbiotic network.

Hereinafter, with reference to the drawings, the text mining method, text mining program, and text mining device of the embodiment of the present invention will be described. The character exploration method of this embodiment is typically executed by a computer. The text exploration program of this embodiment is a program used to implement the text exploration method using a computer. The character exploration device of this embodiment is typically constructed using a computer. The computer running the text mining program functions as a text mining device.

Fig. 1 is a block diagram showing the structure of a character exploration device according to an embodiment of the present invention. The character exploration device 10 shown in FIG. 1 includes an instruction input unit 11, a character data storage unit 12, a word extraction unit 13, a symbiosis matrix generation unit 14, a symbiosis network generation unit 15, and a screen display unit 16. The text mining device 10 generates a symbiosis network based on the text data stored in the text data storage unit 12 as an analysis result of the text data, and displays a screen containing the generated symbiosis network.

The outline of the operation of the text mining device 10 is as follows. To the instruction input unit 11, an instruction from a user (analyzer of text data) is input. The text data storage unit 12 stores one or more text data that can be freely described. The word extraction unit 13 reads the designated text data from the text data storage unit 12, and performs morphological analysis on the read text data, thereby extracting words from the text data. The co-occurrence matrix generation unit 14 generates a co-occurrence matrix for the words extracted by the word extraction unit 13. The symbiosis network generation unit 15 generates a symbiosis network based on the symbiosis matrix generated by the symbiosis matrix generation unit 14. The screen display unit 16 displays a screen including the symbiosis network generated by the symbiosis network generation unit 15.

The user uses the instruction input unit 11 to input instructions to specify the text data of the analysis object, instructions to specify the word of interest, and the like. The word extraction unit 13, the symbiosis network generation unit 15, and the screen display unit 16 perform actions for displaying a screen including the symbiosis network according to instructions from the user. When inputting instructions with designated text data, generate an overall symbiosis network based on the entirety of the designated text data, and display a screen containing the overall symbiosis network. When inputting instructions for specifying the word of interest in the screen containing the overall symbiosis network, a limited symbiosis network based on the sentence containing the word of interest in the specified text data is generated, and the screen containing the limited symbiosis network is displayed.

FIG. 2 is a block diagram showing the structure of a computer functioning as a text exploration device 10. The computer 20 shown in FIG. 2 includes a CPU (Central Processing Unit) 21, a main memory 22, a storage unit 23, an input unit 24, a display unit 25, a communication unit 26, and a recording medium reading unit 27. The main memory 22 uses, for example, DRAM (Dynamic Random Access Memory, dynamic random access memory). The storage unit 23 uses, for example, a hard disk or a solid-state drive. The input unit 24 includes a keyboard 28 or a mouse 29, for example. The display unit 25 uses, for example, a liquid crystal display. The communication unit 26 is an interface line for wired communication or wireless communication. The recording medium reading unit 27 is an interface circuit of the recording medium 30 storing programs and the like. The recording medium 30 uses, for example, CD-ROM (Compact Disc-Read Only Memory), DVD-ROM (Digital Versatile Disc-Read Only Memory), USB (Universal Serial Bus, universal serial bus). Row) non-transitory recording media such as memory.

When the computer 20 executes the text exploration program 31, the storage unit 23 stores the text exploration program 31 and the text data 32. The text mining program 31 and text data 32 can be, for example, those received from a server or other computer using the communication unit 26, or can be read from the recording medium 30 using the recording medium reading unit 27.

When the text mining program 31 is executed, the text mining program 31 and text data 32 are copied and transmitted to the main memory 22. The CPU 21 uses the main memory 22 as a working memory and executes the text exploration program 31 stored in the main memory 22, thereby performing the process of extracting words from the text data 32 and generating a co-occurrence matrix for the extracted words Processing, processing of generating a symbiosis network based on the generated symbiosis matrix, processing of displaying a screen containing the generated symbiosis network, etc. At this time, the computer 20 functions as the character exploration device 10. Furthermore, the configuration of the computer 20 described above is just an example, and any computer can be used to construct the text exploration device 10.

FIG. 3 is a flowchart showing the operation of the text exploration device 10. Before the operation shown in FIG. 3 is performed, the text data storage unit 12 stores more than one text data that can be freely described. Each text data contains multiple sentences. The text mining device 10 processes text data designated by the user among text data stored in the text data storage unit 12.

In FIG. 3, the instruction input unit 11 first receives an instruction to specify text data from the user (step S101). At this time, in addition to receiving instructions for specifying text data, the instruction input unit 11 can also receive instructions for setting the reference value of the co-occurrence matrix (details will be described below), switching between AND processing and OR processing (details will be described below) The instructions for setting the details of the symbiosis network display status, etc. The received instructions are output to each part of the text exploration device 10.

Next, the word extraction unit 13 reads out the designated text data from the text data storage unit 12 (step S102). Next, the word extracting unit 13 extracts words from the read text data by performing morphological analysis on the text data read in step S102 (step S103). At this time, the word extraction unit 13 extracts only the words needed in the subsequent analysis from the read text data. Next, the co-occurrence matrix generation unit 14 uses the text data read out in step S102 to generate a co-occurrence matrix for the words extracted in step S103 (step S104).

FIG. 4 is a diagram showing an example of the co-occurrence matrix generated by the co-occurrence matrix generating unit 14. The element of the co-occurrence matrix is the Jaccard coefficient obtained from the pair of words. Regarding the text data of the analysis object, set the set of sentences containing the word Wa as A, and set the set of sentences containing the word Wb as B. The Jaccard coefficient K(Wa, Wb) of the word pair (Wa, Wb) is provided by the following formula (1).

K(Wa,Wb)=｜A∩B｜/｜A∪B｜…(1)

Among them, in formula (1), the symbol ∩ indicates the operation to find the intersection, the symbol ∪ indicates the operation to find the union, and |S| indicates the number of elements contained in the set S.

In step S104, the co-occurrence matrix generation unit 14 obtains Jaccard coefficients for all pairs of words extracted from the entire text data read in step S102, and generates a co-occurrence matrix with the obtained Jaccard coefficients as elements. The columns and rows of the co-occurrence matrix correspond to the types of words extracted from the entire text data read out in step S102. When extracting n types of words from the whole of the read text data, the co-occurrence matrix generated in step S104 is a symmetric matrix with n columns and n rows with all diagonal elements of 1.

Furthermore, the co-occurrence matrix generation unit 14 may also divide the text data into units other than sentences to obtain the Jaccard coefficient. For example, the co-occurrence matrix generating unit 14 may also set the set of paragraphs containing the word Wa as A, and set the set of paragraphs including the word Wb as B, and obtain the Jaccard coefficient according to formula (1). Moreover, in the case that the sentence contained in the text data has a date, the co-occurrence matrix generating unit 14 may also divide the text data into plural parts including sentences with the same date, and set the set of parts containing the word Wa as A , Set the set of parts containing the word Wb as B, and calculate the Jaccard coefficient according to formula (1). In addition, the co-occurrence matrix generating unit 14 may also generate a co-occurrence matrix including other values (for example, Simpson coefficient or cosine distance, etc.) representing the co-occurrence of words as elements.

Next, the symbiosis network generation unit 15 generates an overall symbiosis network based on the symbiosis matrix generated in step S104 (step S105). Next, the screen display unit 16 displays a screen including the overall symbiosis network generated in step S105 (step S106). FIG. 5 is a diagram showing an example of the window including the overall symbiotic network displayed in step S106. The window 41 shown in FIG. 5 includes an overall symbiosis network 51 and an analysis button 61. The analysis button 61 is provided to instruct the start of analysis.

The symbiosis network generation unit 15 has a reference value of the symbiosis matrix (hereinafter, referred to as V). The reference value V may be a predetermined value, or may be a value set by the user using the instruction input unit 11. In the co-occurrence matrix generated in step S104, when the maximum value of the Jaccard coefficient K(Wa,*) contained in the column corresponding to the word Wa is greater than the reference value V, the co-occurrence network generating unit 15 uses the word Wa The corresponding node (the node recorded as the word Wa) is included in the overall symbiosis network. In addition, in the co-occurrence matrix generated in step S104, when the Jaccard coefficient K(Wa, Wb) of the word pair (Wa, Wb) is greater than the reference value V, the co-occurrence network generating unit 15 will correspond to the word Wa The nodes of and the edges connected to the nodes corresponding to the word Wb are included in the overall symbiosis network.

In the overall symbiosis network 51 shown in FIG. 5, the nodes corresponding to the words with higher occurrence frequency are displayed larger. Alternatively, when displaying a screen that includes a symbiotic network, when the Jaccard coefficient K(Wa, Wb) is large, the edge connecting the node corresponding to the word Wa and the node corresponding to the word Wb is displayed boldly. In addition, the color of the side can be switched according to the Jaccard coefficient, and both the thickness and color of the side can be switched. The symbiosis network is divided into multiple parts that can be reached via edges. When displaying the screen containing the symbiotic network, the multiple nodes contained in each part can also be displayed in the colors assigned to each part. Furthermore, the positions of nodes and edges included in the symbiosis network are meaningless.

Next, the instruction input unit 11 receives an instruction to specify the word of interest from the user (step S111). When step S111 is executed, a screen including the overall symbiotic network is displayed. The user selects the elements of the overall symbiosis network by operating the mouse 29, and inputs instructions for specifying the word of interest. Furthermore, when inputting instructions, the user may use the keyboard 28 instead of the mouse 29, and may also perform operations such as directly touching the display screen. Hereinafter, when step S111 is executed, a screen including the window 41 shown in FIG. 5 is displayed.

6 to 11 are diagrams respectively showing the first to sixth operations of specifying the word of interest in the window 41. In FIG. 6 to FIG. 11, the action instruction frame indicates the sequence of operations, and the white arrow indicates the movement of the mouse cursor 62. The action indicator frame and arrow are not displayed in the actual screen. Hereinafter, the button of clicking (double-clicking) the mouse 29 when the mouse cursor 62 is on an element in the display screen is referred to as "clicking (double-clicking) an element".

As shown in Figure 6, the user first clicks the node (the first click) corresponding to the word designated as the word of interest (here, "open-air bath") in the window 41, and then clicks the analysis button 61 ( Click the second time). With this operation, the word corresponding to the node clicked for the first time is designated as the word of interest. In this way, select a node included in the overall symbiosis network in the screen including the overall symbiosis network, and select the start of analysis, thereby inputting instructions for specifying a word of interest.

As shown in FIG. 7, the user double-clicks the node corresponding to the word designated as the word of interest (here, "open-air bath") in the window 41. With this operation, the word corresponding to the double-clicked node is designated as the word of interest. In this way, by continuing to select one node included in the overall symbiosis network in the screen containing the overall symbiosis network, and input an instruction to specify a word of interest.

As shown in Figure 8, the user first clicks on the node (the first click) corresponding to the word designated as the word of interest (here, "open-air bath") in the window 41, and then clicks and designates as the focus Another word of the word (here, "price") corresponds to the node (the second click), and finally click the analysis button 61 (the last click). Through this operation, two words corresponding to the nodes clicked the first time and the second time are designated as the words of interest. The user can also click p (p is an integer greater than 3) nodes in the window 41, and finally click the analysis button 61. Through this operation, p words corresponding to p nodes are designated as words of interest. In this way, by selecting a plurality of nodes included in the overall symbiosis network in the screen containing the overall symbiosis network, and selecting the start of analysis, input instructions for specifying a plurality of words of interest.

As shown in FIG. 9, the user double-clicks in the window 41 to connect the two nodes corresponding to the two words designated as the word of interest (here, "open-air bath" and "stairway"). In this way, two words corresponding to the two nodes connected to the side of the double-click are designated as the words of interest. In this way, by continuing to select an edge included in the overall symbiosis network in the screen containing the overall symbiosis network, and input instructions for specifying 2 words of interest.

As shown in Figure 10, the user first clicks in the window 41 to connect the two nodes corresponding to the two words designated as the word of interest (here, "open-air bath" and "stairway") (the first time Click), and then click the analysis button 61 (click for the second time). In this way, two words corresponding to the two nodes connected to the edge clicked for the first time are designated as the words of interest. In this way, by selecting one side included in the overall symbiosis network in the screen including the overall symbiosis network, and selecting the start of analysis, input instructions for specifying 2 words of interest.

As shown in Figure 11, the user first clicks in the window 41 the edge that connects the two nodes corresponding to the two words designated as the word of interest (here, "open-air bath" and "stairway") (first time Click on), then click on the edge that connects the 2 nodes corresponding to the other 2 words (here, "price" and "consideration") designated as the word of interest (the second click), and finally click Select the analysis button 61 (the last click). Through this operation, 4 words corresponding to 4 nodes connected to the 2 edges of the first and second clicks are designated as the words of interest. The user can also click on the q (q is an integer greater than 3) side in the window 41, and finally click the analysis button 61. By this operation, 2q words corresponding to 2q nodes connected to q edges are designated as the words of interest. In this way, by selecting a plurality of edges included in the overall symbiosis network in the screen containing the overall symbiosis network, and selecting the start of analysis, input instructions for specifying a plurality of words of interest.

In step S111, the instruction input unit 11, in addition to receiving instructions for specifying the word of interest, it can also receive instructions for setting the reference value of the symbiosis matrix, instructions for switching between AND processing and OR processing, and setting details of the symbiosis network display mode The instructions etc. The received instructions are output to each part of the text exploration device 10.

Next, the word extracting unit 13 extracts the sentence containing the word of interest specified in step S111 from the text data read out in step S102, and obtains the limited text data of the sentence containing the word of interest (step S112).

The word extraction unit 13 has a flag indicating which of the AND processing and the OR processing is performed when a plurality of words of interest are designated. The value of the flag may be a predetermined value, or may be a value set by the user using the instruction input unit 11. In the case where the flag indicates AND processing, the word extraction unit 13 extracts all sentences including the designated plural attention words from the read text data to obtain the limited text data. In the case where the flag indicates OR processing, the word extraction unit 13 extracts a sentence containing any specified word of interest from the read text data to obtain the limited text data.

Next, the word extraction unit 13 extracts words from the limited text data by performing morphological analysis on the limited text data obtained in step S112 (step S113). Next, the co-occurrence matrix generation unit 14 uses the limited character data obtained in step S112 to generate a co-occurrence matrix for the words extracted in step S113 (step S114). Next, the symbiosis network generation unit 15 generates a limited symbiosis network based on the symbiosis matrix generated in step S114 (step S115). Furthermore, between steps S103 to S105 and steps S113 to S115, the processing objects are different, but the processing content is the same.

Generally speaking, the types of words extracted from the limited text data obtained in step S112 are less than the types of words extracted from the text data read out in step S102. The co-occurrence matrix generated in step S114 is different from the co-occurrence matrix generated in step S104. The limited symbiosis network generated in step S115 is different from the overall symbiosis network generated in step S105.

Next, the screen display unit 16 displays a screen including the limited symbiotic network generated in step S115 (step S116). FIGS. 12 and 13 are diagrams showing examples of windows including a limited symbiotic network displayed in step S116. The window 42 shown in FIG. 12 includes a limited symbiosis network 52 when one word of interest (here, "open-air bath") is specified. The window 43 shown in FIG. 13 includes a limited symbiosis network 53 when two attention words (here, "open-air bath" and "bath") are specified.

14 and 15 are diagrams showing examples of the display screen of the character exploration device 10. The screen display unit 16 can display the windows including the overall symbiosis network and the windows including the limited symbiosis network side by side without overlapping, or display the two overlapped. In the screen 71 shown in FIG. 14, the window 41 including the overall symbiosis network 51 and the window 42 including the limited symbiosis network 52 are displayed side by side without overlapping. The user can observe the overall symbiosis network 51 and the limited symbiosis network 52 in the screen 71 at the same time. In the screen 72 shown in FIG. 15, the window 42 including the limited symbiosis network 52 and the window 41 including the overall symbiosis network 51 are overlapped and displayed. The user can switch to observe the overall symbiosis network 51 and the limited symbiosis network 52 in the screen 72.

Next, the instruction input unit 11 receives an instruction from the user (step S121). Next, the text mining device 10 determines whether the instruction received in step S121 is an instruction to specify the word of interest (step S122). In the case of Yes in step S122, the control of the character exploration device 10 proceeds to step S112. In this case, steps S112 to S116 are performed on the word of interest specified in step S121, and a screen containing a limited symbiosis network based on the limited text data including the sentence containing the word of interest specified in step S121 is displayed.

FIG. 16 is a diagram showing an example of the display screen of the character exploration device 10. In the screen 73 shown in Fig. 16, the window 41 including the overall symbiosis network 51 overlaps the window 42 including the limited symbiosis network 52, and the window including the limited symbiosis network 54 when the word "bath" is designated as the word of interest is displayed. 44. The screen 73 is displayed when "open-air bath" is designated as the word of interest in step S111, and "bath" is designated as the word of interest in step S121. On the screen 73, the user can switch to observe the overall symbiosis network 51 and the limited symbiosis network 52, 54.

In the case of No in step S122, the control of the character exploration device 10 proceeds to step S123. In this case, the instruction received in step S121 is, for example, an instruction to move the window, an instruction to set the window to non-display, an instruction to close the window, an instruction to merge windows, etc. The user inputs the instructions by operating the instruction input unit 11 when displaying the screen including the overall symbiosis network and the limited symbiosis network. The screen display unit 16 displays the updated screen according to the instruction received in step S121 (step S123). Then, the control of the character surveying apparatus 10 proceeds to step S121.

Figure 17 is a diagram showing the operation of merging windows. The screen 74 shown in Fig. 17 shows a window 42 containing a limited symbiosis network 52 when "open-air bath" is designated as the word of interest and a window 44 containing a limited symbiosis network 54 when "bath" is designated as the word of interest . The user can observe two limited symbiotic networks 52 and 54 on the screen 74 at the same time.

The hatched arrow shown in FIG. 17 represents the movement of the mouse cursor 62 when the button of the mouse 29 is pressed. The arrow is not displayed in the actual screen. The user performs an operation (drag and drop operation) of tapping the limited symbiosis network 52 in the screen 74 and releasing it in the limited symbiosis network 54. In more detail, the user presses the button of the mouse 29 when the mouse cursor 62 is in the window 42 and keeps the state of pressing the button of the mouse 29 to move the mouse cursor 62 into the window 44 on the mouse cursor 62 Release the button of the mouse 29 when you are in the window 44. With this operation, enter the instructions for merging windows.

Fig. 18 is a diagram showing a display screen after the operation shown in Fig. 17 is performed. In the screen 75 shown in FIG. 18, a window 45 for displaying a plurality of limited symbiotic networks in the form of labels is displayed. In FIG. 18, the label 64 recorded as "open-air bath" is selected, and the limited symbiosis network 52 when "open-air bath" is designated as the word of interest is displayed in the window 45. When the label 63 recorded as "bath" is selected, the limited symbiosis network 54 shown in FIG. 17 is displayed in the window 45.

When the user clicks the close button (x mark) in the window 45, the window 45 is closed. When the user clicks the close button in the label 63, the label 63 is not displayed. When the user clicks the close button in the label 64, the label 64 is not displayed, and the limited symbiosis network 54 is displayed in the window 45.

As shown above, the text mining method of this embodiment includes the steps of extracting words from text data (steps S102, S103, S112, S113), and the steps of generating a co-occurrence matrix for the extracted words (steps S104, S114) , Steps of generating a symbiosis network based on the generated symbiosis matrix (steps S105, S115), and displaying a screen containing the symbiosis network (steps S106, S116). When inputting the instruction of the designated word of interest in the first screen (the screen containing the window 41) of the first symbiosis network (the overall symbiosis network 51) based on the specified text data, the steps to capture the word ( Steps S112, S113) are to extract words from the limited text data that includes the part of the word of interest (sentences containing the word of interest) in the specified text data, and the step of generating a co-occurrence matrix (step S114) is to extract the words Use the limited text data to generate the second symbiosis matrix, the step of generating the symbiosis network (step S115) is the step of generating the second symbiosis network (limited symbiosis network 52~54) based on the second symbiosis matrix, and displaying the screen (step S116) It displays the second screen including the second symbiotic network (including the screens of windows 42~45). In this way, in the text exploration method of this embodiment, when inputting the instruction of the specified word of interest in the first screen of the first symbiosis network including the entirety of the specified text data, the display contains the specified text data The second screen of the second symbiosis network in the part containing the word of interest. Therefore, it is possible to display the screen of the symbiosis network including the specified word of interest with a simple operation.

In addition, by selecting one or more nodes included in the first symbiosis network in the first screen, and selecting the start of analysis, input the instruction to designate the word corresponding to the node as the word of interest (Figure 6, Figure 8) ). In this way, by selecting one or more nodes in the first screen and starting the analysis, you can input instructions for specifying one or more words of interest by simple operations, and display the symbiosis when specifying one or more words of interest. The screen of the network. In addition, by continuing to select a node included in the first symbiosis network in the first screen, input the instruction to designate the word corresponding to the node as the word of interest (Figure 7). In this way, by continuing to select a node in the first screen, an instruction to specify a word of interest can be input by a simple operation, and a screen containing the symbiosis network when a word of interest is specified is displayed.

In addition, by continuing to select an edge included in the first symbiosis network in the first screen, inputting the words corresponding to the two nodes connected to the edge is an instruction to designate the word of interest (Figure 9). In this way, by continuing to select an edge in the first screen, an instruction to specify two words of interest can be input by a simple operation, and a screen containing the symbiosis network when two words of interest are specified is displayed. In addition, by selecting one or more edges included in the first symbiosis network in the first screen, and selecting the start of analysis, input the instruction to designate the words corresponding to the plural nodes connected to the edges as the words of interest ( Figure 10, Figure 11). In this way, by selecting one or more edges in the first screen and starting the analysis, you can input instructions for specifying multiple words of interest by simple operations, and display the screen containing the symbiosis network when specifying multiple words of interest.

In addition, when the merge instruction is entered in the second screen (screen 74) that contains a plurality of second symbiosis networks (limited symbiosis networks 52, 54) (Figure 17), the steps for displaying the screen will be the plural second symbiosis networks The road is displayed as a label (Figure 18). In this way, a plurality of second symbiotic networks can be displayed concisely. In addition, by tapping a second symbiosis network (limited symbiosis network 52) on the second screen and releasing it in another second symbiosis network (limited symbiosis network 54), input the merge instruction. Therefore, it is possible to input a merge instruction with a simple operation, and to display a plurality of second symbiotic networks in a simplified manner.

The restricted text data can also be composed of sentences containing the word of interest in the specified text data. In this case, when inputting instructions for specifying the word of interest, the specified text data can be divided into sentence units to obtain the limited text data, and the screen containing the second symbiosis network based on the obtained limited text data is displayed . The limited text data when specifying a plurality of words of interest can also be composed of all sentences containing the plurality of words of interest in the specified text data. In this case, it is possible to display the screen containing the second symbiosis network when AND processing plural words of interest. The limited text data when specifying a plurality of words of interest can also be composed of sentences containing any of the plurality of words of interest in the specified text data. In this case, it is possible to display the screen containing the second symbiosis network when OR processing a plurality of words of interest. In addition, the step of generating a co-occurrence matrix is to generate a co-occurrence matrix with Jaccard coefficients as elements. Therefore, the symbiosis of words contained in the text data can be better analyzed.

The character exploration device 10 and the character exploration program 31 of this embodiment have the same characteristics as the above-mentioned character exploration method, and exert the same effects. According to the text mining method, text mining device 10, and text mining program 31 of the present embodiment, the screen of the symbiosis network including the specified word of interest can be displayed with a simple operation.

The present invention has been described in detail above, but the above description is illustrative in all aspects and not restrictive. It should be understood that many other changes or modifications can be made without departing from the scope of the present invention.

Claims (20)

A text exploration method, which displays a screen containing the analysis results of text data and is executed by a computer; it is characterized in that it has the following steps: a step of extracting words from text data; a step of generating a co-occurrence matrix for the words; The step of generating a symbiosis network based on the aforementioned symbiosis matrix; and the step of displaying a screen containing the aforementioned symbiosis network; entering the specified word of interest in the first screen including the first symbiosis network based on the entirety of the specified text data When instructed, the step of extracting the above word is to extract the word from the limited text data that includes the part of the word of interest in the specified text data, and the step of generating the co-occurrence matrix is to generate the above word using the limited text data In the second symbiosis matrix, the step of generating the symbiosis network is based on the second symbiosis matrix to generate a second symbiosis network, and the step of displaying the screen is to display the second screen including the second symbiosis network. Such as the text exploration method of claim 1, wherein, by selecting one or more nodes included in the first symbiosis network in the first screen, and selecting the start of analysis, input the word corresponding to the node Designated as an indication of the above words of interest. Such as the text exploration method of claim 1, wherein, by continuing to select a node included in the first symbiosis network in the first screen, and inputting the instruction to designate the word corresponding to the node as the word of interest . Such as the text exploration method of claim 1, wherein, by continuing to select an edge included in the first symbiosis network in the first screen, and inputting the word corresponding to the two nodes connected to the edge is designated as Instructions for the words of interest above. Such as the text exploration method of claim 1, wherein, by selecting in the first screen above One or more edges included in the first symbiosis network, select the analysis to start, and input the words corresponding to the plurality of nodes connected to the edges as instructions to designate the words of interest. For example, the text exploration method of claim 1, wherein, when a merge instruction is entered in a second screen containing a plurality of second symbiosis networks, the step of displaying the above screen is to display the plurality of second symbiosis networks in the form of labels. For example, the text exploration method of claim 6, wherein, by clicking on a second symbiosis network in the second screen and releasing it in another second symbiosis network, the merge instruction is input. Such as the text exploration method of claim 1, wherein the above-mentioned limited text data includes the sentence containing the above-mentioned word of interest in the above-mentioned designated text data. Such as the text exploration method of claim 8, wherein the above-mentioned limited text data when specifying a plurality of words of interest include all sentences in the specified text data that contain the plurality of words of interest. For example, the text exploration method of claim 8, wherein the above-mentioned limited text data when a plurality of words of interest are designated includes a sentence containing any one of the above-mentioned plural words of interest in the designated text data. Such as the text exploration method of claim 1, wherein the step of generating the above-mentioned co-occurrence matrix is to generate a co-occurrence matrix with Jaccard coefficient as an element. A recording medium recorded with a text exploration program, the text exploration program is used to display the screen containing the analysis result of the text data; it is characterized in that the CPU uses the memory to make the computer execute the following steps: extract words from the text data Steps; Steps to generate a co-occurrence matrix for the above words; The step of generating a symbiosis network based on the aforementioned symbiosis matrix; and the step of displaying a screen containing the aforementioned symbiosis network; entering the specified word of interest in the first screen including the first symbiosis network based on the entirety of the specified text data When instructed, the step of extracting the above word is to extract the word from the limited text data that includes the part of the word of interest in the specified text data, and the step of generating the co-occurrence matrix is to generate the above word using the limited text data In the second symbiosis matrix, the step of generating the symbiosis network is based on the second symbiosis matrix to generate a second symbiosis network, and the step of displaying the screen is to display the second screen including the second symbiosis network. For example, the recording medium of the text exploration program recorded in the request item 12, in which, by selecting one or more nodes included in the first symbiosis network in the first screen, and selecting the start of analysis, the input will be The word corresponding to the aforementioned node is designated as an indication of the aforementioned word of interest. For example, the recording medium of the text exploration program recorded in the request item 12, wherein, by continuing to select a node included in the first symbiotic network in the first screen, and inputting the word corresponding to the node Instructions for the words of interest above. For example, the recording medium of the text exploration program recorded in the request item 12, in which, by continuing to select one edge included in the first symbiosis network in the first screen above, the input will be connected to the two nodes connected to the above edge The corresponding word is designated as an indication of the above-mentioned word of interest. If the recording medium of the request item 12 is recorded with a text exploration program, the input will be connected to the input by selecting one or more edges included in the first symbiosis network in the first screen and selecting the start of analysis The words corresponding to the plural nodes of the aforementioned edges are designated as the indications of the aforementioned words of interest. For example, if the request item 12 records the recording medium of the text exploration program, in which, when the merge instruction is entered in the second screen containing a plurality of second symbiosis networks, the step of displaying the above screen is to convert the plurality of second symbiosis networks The road is displayed as a label. For example, the recording medium of the text exploration program recorded in the request item 17, wherein, by clicking on a second symbiosis network in the second screen and releasing it in another second symbiosis network, input the merge instruction. A text exploration device that displays a screen containing the analysis result of text data; it is characterized in that it includes: a word extraction unit that extracts words from text data; a co-occurrence matrix generation unit that generates a co-occurrence matrix for the words ; The symbiosis network generation part, which generates the symbiosis network based on the above-mentioned symbiosis matrix; and the screen display part, which displays the screen containing the above-mentioned symbiosis network; in the first symbiosis network that contains the whole based on the specified text data When an instruction to specify a word of interest is input in the first screen, the word extraction unit extracts the word from the limited text data that includes the portion of the word of interest in the specified text data, and the co-occurrence matrix generation unit uses the word The text data is limited to generate a second symbiosis matrix, the symbiosis network generation unit generates a second symbiosis network based on the second symbiosis matrix, and the screen display unit displays a second screen including the second symbiosis network. For example, in the text exploration device of claim 19, when a merge instruction is input in a second screen containing a plurality of second symbiosis networks, the screen display unit displays the plurality of second symbiosis networks in label form.

Images