TW201807597A - Text mining method, text mining program, and text mining apparatus - Google Patents

Abstract

In text analysis steps (S109-S110), hierarchical cluster analysis is carried out for words extracted from inputted text data. In a screen generation step (S111), m clusters are calculated from the analysis result of the text analysis steps on the basis of m groups and the maximum number n of data within each of the groups, and screen data for displaying, on a screen, a group which includes not more than n words in the clusters is generated. In an analysis result display step (S112), the screen is displayed on the basis of the generated screen data. Thus, the result of the hierarchical cluster analysis is displayed on the screen so as to be intuitively understood by a user.

Description

Text exploration method, computer-readable recording medium and text exploration device recorded with text exploration program

The present invention relates to text exploration, and in particular, to a text exploration method, a text exploration program, and a text exploration device that display analysis results of text data on a screen.

In recent years, text exploration that analyzes a large amount of text data recorded in free form and obtains useful information from the analysis results has attracted attention. In text exploration, for example, words are extracted from the text data of the analysis object, and information is obtained by analyzing the appearance frequency and appearance trend of the words.

In the following, a text exploration device that performs hierarchical cluster analysis on the words extracted from the text data and displays the analysis results on the screen is discussed. In hierarchical cluster analysis, clusters containing words with a high degree of similarity are created hierarchically based on the similarity between the words. In general, the results of hierarchical cluster analysis are provided to users (analysts) using a dendrogram shown in FIG. 15.

Related to the invention of this case, a clustering device is described in Patent Document 1, which has a hierarchical tree structure for constructing a tree map, exploring the tree map, and generating a specific index from the lower layer to the upper layer and storing it in a storage means. A citation device is described in Citation 2 which includes: a distance matrix calculation means that calculates the distance between keywords, generates a distance matrix data that can explore the distance between keywords and the keywords, and stores them in a storage means; and grouping Means, which uses a distance matrix to hierarchically classify keywords The clusters are stored in the storage means as a bottom-up index of the constructed tree map that can be explored from the lower layer to the upper layer.

[Prior technical literature] [Patent Literature]

[Patent Document 1] Japanese Patent Laid-Open No. 2011-216021

[Patent Document 2] Japanese Patent Laid-Open No. 2012-150539

The conventional text exploration device uses a tree diagram to display the results of hierarchical cluster analysis on the screen. However, such a text exploration device has a problem that a user cannot intuitively understand the analysis result. For example, in the analysis result shown in FIG. 15, when the user sets the number of clusters to 4, as shown in FIG. 16, a cutting line is set on the tree diagram. However, the user cannot intuitively recognize the words contained in each cluster only by seeing such a tree diagram. In addition, when the number of words is large and the number of clusters is changed, the user cannot intuitively grasp how the words included in each cluster change.

In addition, because the tree diagram does not record the frequency of occurrence of words, the user cannot know which word is more important. In addition, when the text data to be analyzed is time-series data with information such as year, month, day, time, etc., the user may be expected to know the temporal change of the analysis result. However, in the conventional text exploration device, the above expectations of users cannot be met.

Therefore, an object of the present invention is to provide a text exploration method, a text exploration program, and a text exploration device that display the results of hierarchical cluster analysis on a screen in a manner that users can intuitively understand.

The first aspect of the present invention is a text exploration method, which displays the analysis result of text data on the screen, and is characterized by having a text analysis step for the single word (monolingual) extracted from the input text data. (Ie, word, word, vocabulary) for hierarchical cluster analysis; a screen generation step that generates screen data based on the analysis results of the above text analysis step; and an analysis result display step that displays a screen based on the screen data; the screen generation Steps: According to the number of groups and the maximum number of data in the group, the cluster of the number of groups is obtained from the above analysis result, and a group including the single word included in the cluster below the maximum number of data is generated and displayed in the group. Picture data of the picture.

The second aspect of the present invention is characterized in that, in the first aspect of the present invention, the words included in the group are selected from the words included in the cluster corresponding to the group in the order of high occurrence frequency .

The third aspect of the present invention is characterized in that, in the second aspect of the present invention, the group has a total size corresponding to the appearance frequency of the words included in the cluster corresponding to the group in the screen.

A fourth aspect of the present invention is characterized in that, in the third aspect of the present invention, the words included in the group have a size corresponding to the appearance frequency of the words in the screen.

A fifth aspect of the present invention is characterized in that, in the first aspect of the present invention, there is further provided an instruction input step for inputting instructions from a user, any one of the above-mentioned text analysis step and the above-mentioned screen generation step, It is executed according to the instruction input in the instruction input step.

A sixth aspect of the present invention is characterized in that in the fifth aspect of the present invention, the instruction input step receives the setting instruction of the number of groups, and the screen generation step is based on the number of groups set in the instruction input step, To generate the above screen data.

A seventh aspect of the present invention is characterized in that, in the fifth aspect of the present invention, the instruction input step receives the setting instruction of the maximum number of data, and the screen generating step is based on the maximum number of data set in the instruction input step. To generate the above screen data.

An eighth aspect of the present invention is characterized in that, in the fifth aspect of the present invention, the instruction input step receives a setting instruction during an analysis object, and the character analysis step analyzes the character data that is set in the instruction input step. The words included in the text data during the target period were analyzed by the above-mentioned hierarchical cluster analysis.

The ninth aspect of the present invention is characterized in that, in the fifth aspect of the present invention, the instruction input step receives a setting instruction for analysis purpose, and the text analysis step extracts from the text data corresponding to the information input in the instruction input step. Set the type of words for the analysis purpose to perform the above-mentioned hierarchical cluster analysis.

The tenth aspect of the present invention is characterized in that, in the fifth aspect of the present invention, the instruction input step receives a single-word exclusion instruction, and the character analysis step excludes the single word indicated by the instruction input step, and performs the hierarchical structure Cluster analysis.

An eleventh aspect of the present invention is the fifth aspect of the present invention. In the above, the instruction input step receives a synonym registration instruction, and the character analysis step treats the plurality of words indicated in the instruction input step as the same word, and performs the hierarchical cluster analysis.

The twelfth aspect of the present invention is characterized in that, in the fifth aspect of the present invention, the instruction input step receives a compound word registration instruction, and the character analysis step combines a plurality of words indicated in the instruction input step into one word And perform the above-mentioned hierarchical cluster analysis.

The thirteenth aspect of the present invention is characterized in that, in the first aspect of the present invention, the screen generation step generates screen data, the screen data is used to display an analysis result screen including the group, and is used to set the above. Analysis setting screen of analysis result screen.

A fourteenth aspect of the present invention is a computer-readable recording medium that records a text exploration program that displays the analysis results of text data on the screen. It is characterized in that the CPU (central processing unit) uses the memory to make the computer execute the following Steps: a text analysis step, which performs hierarchical cluster analysis on the words extracted from the input text data; a screen generation step, which generates screen data based on the analysis results of the above text analysis step; and a display step of analysis results, It displays the screen according to the above screen data; the above screen generating step obtains the cluster of the above group number from the analysis result according to the number of groups and the maximum number of data in the group, and generates a cluster to include the most data The groups of words included in the above-mentioned clusters are displayed in the screen data of the screen.

A feature of the fifteenth aspect of the present invention is that, in the fourteenth aspect of the present invention, the words included in the group are selected from the words included in the cluster corresponding to the group in the order of high occurrence frequency .

The sixteenth aspect of the present invention is characterized in that, in the fifteenth aspect of the present invention, the group has a total size corresponding to the frequency of occurrence of the words included in the cluster corresponding to the group in the screen.

A seventeenth aspect of the present invention is characterized in that, in the sixteenth aspect of the present invention, the words included in the group have a size corresponding to the appearance frequency of the words in the screen.

The eighteenth aspect of the present invention is characterized in that, in the fourteenth aspect of the present invention, the computer is further caused to execute any one of an instruction input step for inputting instructions from a user, the text analysis step, and the screen generation step. It is executed according to the instruction input in the instruction input step.

The nineteenth aspect of the present invention is characterized in that, in the fourteenth aspect of the present invention, the screen generation step generates screen data, the screen data is used to display the analysis result screen including the group, and to set the analysis The analysis setting screen of the result screen.

The twentieth aspect of the present invention is a text exploration device that displays the analysis results of text data on the screen. It is characterized by having a text analysis unit that hierarchically extracts the words extracted from the input text data. Cluster analysis; screen generation unit that generates screen data based on the analysis results of the text analysis unit; and analysis result display unit that displays screens based on the screen data; the screen generation unit based on the number of groups and within the group The cluster with the largest number of data is obtained from the above analysis result, and a cluster containing the words included in the cluster below the cluster with the largest number of data is generated and displayed on the screen.

The twenty-first aspect of the present invention is characterized in that, in the twentieth aspect of the present invention, the words included in the group are selected from the words included in the cluster corresponding to the group according to the order of high occurrence frequency .

The twenty-second aspect of the present invention is characterized in that, in the twenty-first aspect of the present invention, the group has a total size corresponding to the appearance frequency of the words included in the cluster corresponding to the group in the screen.

The twenty-third aspect of the present invention is characterized in that, in the twenty-second aspect of the present invention, the words included in the group have a size corresponding to the appearance frequency of the words in the screen.

The twenty-fourth aspect of the present invention is characterized in that, in the twentieth aspect of the present invention, It further includes an instruction input unit for inputting an instruction from a user, and either of the character analysis unit and the screen generation unit operates based on the instruction input in the instruction input unit.

The twenty-fifth aspect of the present invention is characterized in that, in the twentieth aspect of the present invention, the screen generation unit generates screen data, the screen data is used to display an analysis result screen including the group, and to set the analysis The analysis setting screen of the result screen.

According to the first, fourteenth, or twentieth aspects of the present invention, based on the result of performing hierarchical cluster analysis on the words included in the text data, a group including the words included in the cluster is displayed on the screen. The number of words in a group is limited to the maximum number of data. Therefore, users can intuitively understand the results of hierarchical cluster analysis when they see the screen.

According to the second, the fifteenth, or the twenty-first aspect of the present invention, within the group, a word having a high frequency among the words included in the cluster is displayed. Therefore, the user can easily recognize the frequently occurring words contained in each cluster.

With the third, sixteenth, or twenty-second aspect of the present invention, the group has a total size corresponding to the appearance frequency of the words included in the cluster in the screen. Therefore, the user can easily recognize a large cluster of the total occurrence frequency of the single word.

With the 4th, 17th, or 23rd aspect of the present invention, the single word has a size corresponding to the frequency of the single word in the screen. Therefore, the user can easily recognize words that appear frequently.

According to the fifth, eighteenth or twenty-fourth aspect of the present invention, The user instructs to switch the display of the results of hierarchical cluster analysis.

According to the sixth aspect of the present invention, the number of groups (the number of clusters) displayed on the screen can be switched in accordance with an instruction from the user.

According to the seventh aspect of the present invention, the upper limit of the number of words included in the group can be switched in response to an instruction from the user.

According to the eighth aspect of the present invention, the result of performing hierarchical cluster analysis on the words included in the text data within the analysis target period instructed by the user is displayed on the screen. Therefore, the user can easily recognize the temporal change of the results of the hierarchical cluster analysis.

According to the ninth aspect of the present invention, corresponding to the analysis purpose instructed by the user, the type of the single word of the analysis target can be switched and the result of the hierarchical cluster analysis can be displayed on the screen.

According to the tenth aspect of the present invention, the words instructed by the user can be excluded, and the result of the hierarchical cluster analysis can be displayed on the screen.

According to the eleventh aspect of the present invention, the plurality of words instructed by the user can be regarded as the same word, and the result after the hierarchical cluster analysis is displayed on the screen.

According to the twelfth aspect of the present invention, the plurality of words instructed by the user can be combined into one word, and the result of the hierarchical cluster analysis is displayed on the screen.

According to the 13th, 19th or 25th aspect of the present invention, the analysis result screen and the analysis setting screen are displayed. Therefore, the user can easily switch the display state of the result after performing the hierarchical cluster analysis using the analysis setting screen.

5‧‧‧Text

10‧‧‧Text exploration device

11‧‧‧Instruction input section

12‧‧‧Text Analysis Department

13‧‧‧Screen generation department

14‧‧‧ Analysis result display section

20‧‧‧Computer

21‧‧‧CPU

22‧‧‧Main memory

23‧‧‧Storage Department

24‧‧‧ Input Department

25‧‧‧Display

26‧‧‧ Ministry of Communications

27‧‧‧Recording medium reading section

28‧‧‧ keyboard

29‧‧‧Mouse

30‧‧‧Recording media

31‧‧‧text exploration program

40‧‧‧display

41, 61 ~ 68‧‧‧ Analysis result screen

42‧‧‧ Analysis Setting Screen

51‧‧‧Data Designation Screen

52‧‧‧ Purpose Designation Screen

53‧‧‧Synonym list selection screen

54‧‧‧ Compound word list selection screen

m‧‧‧groups (number of clusters)

n‧‧‧The maximum number of data in the group

W1 ~ W6‧‧‧Word

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

FIG. 2 is a block diagram showing the configuration of a computer functioning as the text exploration device shown in FIG. 1. FIG.

FIG. 3 is a diagram showing a display screen of the text exploration device shown in FIG. 1. FIG.

FIG. 4 is a flowchart showing the operation of the text exploration device shown in FIG. 1. FIG.

FIG. 5 is a flowchart of a screen data generating process of the text exploration device shown in FIG. 1. FIG.

FIG. 6 is a diagram showing a data specifying screen of the text exploration device shown in FIG. 1. FIG.

FIG. 7 is a diagram showing an example of text data input to the text exploration device shown in FIG. 1. FIG.

FIG. 8 is a diagram showing a purpose designation screen of the text exploration device shown in FIG. 1. FIG.

FIG. 9 is a diagram showing a synonym list selection screen of the text exploration device shown in FIG. 1. FIG.

FIG. 10 is a diagram showing a compound word list selection screen of the text exploration device shown in FIG. 1. FIG.

FIG. 11A is a diagram showing an analysis result screen before the analysis target period is set in the text exploration device shown in FIG. 1. FIG.

FIG. 11B is a diagram showing an analysis result screen after the analysis target period is set in the text exploration device shown in FIG. 1.

FIG. 12A is a diagram showing an analysis result screen before a single character is excluded in the text exploration device shown in FIG. 1. FIG.

FIG. 12B is a diagram showing an analysis result screen after single words are excluded in the text exploration device shown in FIG. 1.

FIG. 13A is a diagram showing an analysis result screen before synonyms are registered in the text exploration device shown in FIG. 1. FIG.

FIG. 13B is a diagram showing an analysis result screen after synonyms are registered in the text exploration device shown in FIG. 1.

FIG. 14A is a diagram showing an analysis result screen before compound word registration is performed in the text exploration device shown in FIG. 1. FIG.

FIG. 14B is a diagram showing an analysis result screen after compound word registration is performed in the text exploration device shown in FIG. 1. FIG.

FIG. 15 is a diagram showing an example of a tree diagram.

FIG. 16 is a diagram showing a case where the number of clusters is set to the tree diagram shown in FIG. 15.

FIG. 17 is a diagram showing words appearing in the drawing and its description.

Hereinafter, a character exploration method, a character exploration program, and a character exploration device according to an embodiment of the present invention will be described with reference to the drawings. The text exploration method of this embodiment is usually executed by a computer. The text exploration program of this embodiment is a program for implementing a text exploration method using a computer. The character exploration device of this embodiment is usually constructed using a computer. The computer executing the text exploration program functions as a text exploration device.

FIG. 1 is a block diagram showing the structure of a text exploration device according to an embodiment of the present invention. The character exploration apparatus 10 shown in FIG. 1 includes an instruction input unit 11, a character analysis unit 12, a screen generation unit 13, and an analysis result display unit 14. The text data 5 of the analysis target is input into the text exploration device 10. The character exploration device 10 performs hierarchical cluster analysis on the words extracted from the input character data 5 and displays the analysis results on the screen.

The outline of the operation of the character exploration apparatus 10 is as follows. An instruction from the user is input to the instruction input unit 11. The character analysis unit 12 extracts characters from the input character data 5 and performs hierarchical cluster analysis on the extracted characters. The screen generation unit 13 generates screen data based on the analysis result of the character analysis unit 12. The analysis result display unit 14 displays a screen based on the screen data generated by the screen generating unit 13.

The instructions from the user input to the instruction input section 11 include the setting of the number of groups, the setting of the maximum number of data in the group, the setting of the analysis target period, excluding single words, synonyms, compound words, and the like. When the text data 5 is time-series data with information such as year, month, day, time, etc., the text analysis unit 12 analyzes the text data of the input text data 5 within the analysis target period set by the instruction input unit 11. Included words for hierarchical cluster analysis.

The screen generating unit 13 generates screen data according to the number of groups and the maximum number of data in the group (the details will be described later). When a user inputs a new instruction, after the instructed processing is performed, the screen generating unit 13 generates new screen data, and the analysis result display unit 14 displays a new screen. In this way, the text exploration device 10 switches the analysis state of the text data 5 and the display state of the analysis result in response to an instruction from the user.

FIG. 2 is a block diagram showing a configuration of a computer functioning as the text exploration device 10. As shown in FIG. The computer 20 shown in FIG. 2 includes a CPU (Central Processing Unit) 21, a main memory 22, a storage section 23, an input section 24, a display section 25, a communication section 26, and a recording medium reading section 27. . The main memory 22 is, for example, a DRAM (Dynamic Random Access Memory). The storage unit 23 uses, for example, a hard disk (Hard Disk) or a solid state drive (Solid State Drive). The input unit 24 includes, for example, a keyboard 28 and a mouse (Mouse) 29. The display unit 25 is, for example, a liquid crystal display. The communication section 26 is an interface circuit for wired communication or wireless communication. The recording medium reading unit 27 is an interface circuit of the recording medium 30 in which a program or the like is stored. The recording medium 30 is, for example, a CD-ROM (Compact Disc Read-Only Memory), a DVD-ROM (Digital Versatile Disc Read-Only Memory), or a USB (Universal Serial Bus; (Universal Serial Bus) 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 5. The text exploration program 31 and the text data 5 may be, for example, those who use the communication unit 26 from a server or other computer receivers, or those who use the recording medium reading unit 27 from the recording medium 30.

When the text exploration program 31 is executed, the text exploration program 31 and the text data 5 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 to process the character data 5 stored in the main memory 22. At this time, the computer 20 functions as the character exploration device 10. In addition, the configuration of the computer 20 described above is merely an example, and an arbitrary computer may be used to configure the text exploration device 10.

Hereinafter, the text data 5 is assumed to be Japanese data including Japanese characters. FIG. 17 is a diagram showing words appearing in the diagram and its description. Words (Japanese words) and meanings of the words are described in each column of FIG. 17. When referring to a Japanese word in the following description, the meaning of the word may be described in parentheses after the word. Furthermore, the text data 5 may be data in any language.

FIG. 3 is a diagram showing a display screen of the text exploration device 10. The display screen 40 shown in FIG. 3 includes an analysis result screen 41 and an analysis setting screen 42. The analysis result of the character analysis unit 12 is displayed on the analysis result screen 41. A GUI (Graphical User Interface) element is displayed on the analysis setting screen 42, and the GUI element is used to set the analysis mode of the character analysis unit 12 and the characteristics of the screen data generated by the screen generation unit 13.

If the number of clusters is set for the result of hierarchical cluster analysis, the words included in each cluster are determined. When the results of hierarchical cluster analysis of the words extracted from the text data 5 are displayed on the screen, the text exploration device 10 displays the groups corresponding to the clusters as shown in FIG. 3 instead of the tree diagram .

In the following description, the clusters displayed on the screen are also called groups. The user uses the instruction input unit 11 to specify the number of groups (the number of clusters) and the maximum number of data in the group (the upper limit of the number of single words included in the group). Hereinafter, the former is set to m and the latter is set to n.

In the character exploration apparatus 10, the single characters included in the character data 5 are classified into m clusters, and each cluster includes one or more single characters. M groups are displayed on the analysis result screen 41, and single words are displayed inside each group. The group system is displayed using a cloud-like graphic, and the individual words contained in the group are displayed inside the ellipse area. The number of words contained in each group is limited to n or less. For example, when the cluster at the time of n = 5 contains 10 words, in the analysis result screen 41, 5 words are displayed inside the group.

On the analysis setting screen 42, a first slider bar for setting the number of groups m and two first buttons (marked with a mark "+" or "-") are displayed. A second slide bar and two second buttons, and four boxes and two third buttons (the ones marked with a left arrow or a right arrow) for setting an analysis target period.

By operating the mouse 29, the user moves the scroll block of the first slider to the left or right or presses the first button to indicate the group number m. The number of groups m increases when the first button marked with a "+" is pressed, and decreases when the first button marked with a "-" is pressed. The initial value of the number of groups m is set, for example, as the square root of the type of a single word included in the analysis result of the character analysis unit 12, or an integer close to the square root. For example, when the analysis result of the character analysis unit 12 includes 16 types of single words, the initial value of the group number m is set to 4.

By operating the mouse 29, the user moves the scroll block of the second slider to the left or right or presses the second button to indicate the maximum number of data n in the group. The maximum number of data n in the group will increase or decrease when the second button is pressed. The initial value of the maximum number of data n in the group is set to 5, for example.

When the text data 5 is time-series data, the user uses the keyboard 28 or the mouse 29 to designate the year, month, day, and time or press the third button to indicate the analysis target period. During the analysis period, when the third button marked with a left arrow is pressed, it moves toward the past by a predetermined amount (for example, one month), and when the third button marked with a right arrow is pressed, it moves in the opposite direction. Both quantitative. The initial value of the analysis target period is set, for example, as a period from the oldest time to the latest time of the text data 5. When the text data 5 is not time-series data, the user cannot specify the analysis target period.

One or more and m or less groups are displayed on the analysis result screen 41, and one or more and n or less words are displayed inside each group. Each group in the screen, the larger the total frequency of the occurrence of the words contained in the corresponding cluster, the larger it will be displayed. When the number of words included in the cluster exceeds n, n words with high frequency are displayed in the group. Words included in groups and ellipses containing them The circle area, the higher the frequency of single words in the screen, the larger the size will be displayed. Each group is marked with a name. The group name uses the word that appears most frequently among the words contained in the cluster. The name of the group is displayed on the underline of the group. When a single character cannot be displayed inside the elliptical region, the symbol "..." is displayed instead of the single character.

On the analysis result screen 41, a third slide bar for specifying a zoom ratio and two fourth buttons (those marked with "+" or "-") are displayed. The user sets the zoom ratio by operating the mouse 29 to move the scroll block of the third slider to the left or right or pressing the fourth button. On the analysis result screen 41, groups containing single words are displayed enlarged or reduced in accordance with a set zoom ratio. The initial value of the zoom ratio is set to 100%. On the analysis result screen 41 in the initial state, all groups are displayed.

When the user changes the number of groups m, the maximum number of data n in the group, or the analysis target period in the analysis setting screen 42, the content of the analysis result screen 41 changes in accordance with these. When the user instructs the exclusion of a single word, the registration of a synonym, or the registration of a compound word on the analysis result screen 41, the content of the analysis result screen 41 also changes according to these.

When performing hierarchical cluster analysis on the words extracted from the text data 5, the text exploration device 10 refers to the list of excluded words that stores the words that should be excluded, the list of synonyms that stores the words that should be treated as synonyms, and the A list of compound words that should be treated as compound words. A plurality of words having the same meaning (or approximately the same meaning) and one word representing these words are associated and stored in the synonyms list. If connected, a plurality of words that become a compound word are associated with the compound words that link these words and are stored in the compound word list. For example, "daigakusei (university student)" and "gakusei (student)" and "daigakusei" representing both Correspondence is stored in the synonyms list. For example, "nintai (endurance)" and "tsuyoi (strong)" are associated with "nintaizuyoi (endurance)" and are stored in the compound word list. The text exploration apparatus 10 may have a plurality of synonyms and a plurality of compound words.

FIG. 4 is a flowchart showing the operation of the text exploration device 10. FIG. 5 is a flowchart showing the details of the screen data generation processing (step S111 shown in FIG. 4) of the text exploration device 10. The input unit 24 and the CPU 21 executing step S113 function as an instruction input unit 11. The CPU 21 that executes steps S109 to S110 functions as the character analysis unit 12. The CPU 21 that executes step S111 functions as the screen generating section 13. The display section 25 and the CPU 21 executing step S112 function as the analysis result display section 14. Hereinafter, the operation of the character exploration apparatus 10 will be described with reference to FIGS. 4 and 5.

First, the CPU 21 causes the display unit 25 to display the data specifying screen 51 shown in FIG. 6 (step S101). A box for specifying a file name and a box for specifying a folder name are displayed on the data specifying screen 51. The user specifies the text data 5 to be analyzed by specifying a file name or a folder name in the data specifying screen 51. The text data 5 may be stored in a storage section 23 such as a hard disk, or in a server or other computer connected to the communication section 26.

Then, the CPU 21 transmits the character data 5 specified by the use data designation screen 51 to the main memory 22. Thereby, the character data 5 is input to the character exploration apparatus 10 (step S102). FIG. 7 is a diagram showing an example of the text data 5. FIG. The textual data shown in Figure 7 is the data of reports produced by college students, and is time-series data with the information of year, month, and day. The text information shown in Figure 7 is, in order from above, "About the relationship between college students and society in this course ...", "General college students play before leaving for society after graduation. Or "...", "our students need to know that they paid expensive tuition for studying ...", and "student life is a precious time for self-confidence. And ..." The type of the character data 5 analyzed by the character exploration device 10 is arbitrary.

Next, the CPU 21 causes the display unit 25 to display the purpose designation screen 52 shown in FIG. 8 (step S103). On the purpose designation screen 52, three option buttons (Radio Buttons) corresponding to content, characteristics, and evaluation are displayed. The user selects an analysis purpose from the content, features, and evaluation by operating the mouse 29 and pressing any option button. Then, the CPU 21 receives the analysis purpose specified by the use purpose designation screen 52. Thereby, the analysis purpose is input to the character exploration apparatus 10 (step S104).

Next, the CPU 21 causes the display unit 25 to display the synonyms list selection screen 53 shown in FIG. 9 (step S105). The synonyms list selection screen 53 displays the names of the synonyms list included in the text exploration device 10 and the synonyms that are registered in the synonyms list. The user selects any synonym list in the synonym list selection screen 53 by operating the mouse 29 to specify a synonym list to be used. Thereby, a synonym list is selected in the character exploration apparatus 10 (step S106).

Next, the CPU 21 causes the display unit 25 to display the compound word list selection screen 54 shown in FIG. 10 (step S107). The compound word list selection screen 54 displays the name of the compound word list included in the text exploration device 10 and the compound words registered in each compound word list. The user selects any compound word list in the compound word list selection screen 54 by operating the mouse 29 to specify a compound word list to be used. Thereby, the compound word list is selected in the character exploration apparatus 10 (step S108).

Next, the CPU 21 considers the list of excluded words, the list of synonyms, and the list of compound words, and the text data belonging to the analysis target period from the text data 5 input in step S102, extracts the type corresponding to the analysis purpose specified in step S104. Word (step S109). When the analysis purpose is "content", the CPU 21 extracts nouns, proper nouns, place names, and person names from the text data 5. When the analysis purpose of the CPU 21 is "feature", the CPU 21 extracts nouns, proper nouns, (SA) The use of nouns and verbs in a declension. When the analysis purpose is "evaluation", the CPU 21 extracts adjectives, adjective verbs, and interjections from the text data 5. In addition, the text exploration device 10 can also support analysis purposes other than the foregoing three. In addition, the CPU 21 can also retrieve different types of words according to the analysis purposes.

When the text data 5 is time-series data, when the CPU 21 executes step S109, it only extracts words from the text data included in the analysis target period indicated by the user in the text data 5. When the word W1 is stored in the excluded word list, the CPU 21 completely ignores the word W1 included in the text data 5 when executing step S109. In addition, when the single word W2 and the single word W3 and the single word W2 representing both are associated and stored in the selected synonym list, the CPU 21 will execute all the single words W3 included in the text data 5 when executing step S109. Treated as single word W2. In addition, when the single word W4 and the single word W5 and the connected single word W6 are associated and stored in the selected compound word list, the CPU 21 executes step S109 to connect the single words included in the text data 5 W4 and single word W5 are all treated as single word W6.

Next, the CPU 21 performs hierarchical cluster analysis on the words retrieved in step S109 (step S110). In step S110, the CPU 21 obtains the similarity between the two words according to the distance between the two words in the text data 5 (how far apart the two words are from each other). The CPU 21 performs hierarchical clustering using a predetermined method (for example, the shortest distance method, the longest distance method, the group average method, the decimal method, the Ward's Method, etc.) according to the similarity between the obtained words. analysis. also, The CPU 21 obtains the appearance frequency of each word in step S110.

Next, the CPU 21 generates screen data for displaying the analysis result based on the result of the hierarchical cluster analysis obtained in step S110 (step S111). The CPU 21 performs processing shown in FIG. 5 in step S111.

The CPU 21 sets the number of groups to m and sets the maximum number of data in the group to n (step S201). Next, the CPU 21 sets the number of clusters to m as a result of the hierarchical cluster analysis to obtain m clusters (step S202). Next, the CPU 21 obtains the total of the appearance frequencies of the words included in the clusters for each cluster (step S203). Next, the CPU 21 determines the display size of each group based on the total of the appearance frequencies obtained in step S203 (step S204). In step S204, the larger the total frequency of occurrences of the words included in the cluster, the larger the display size of the group is determined.

Next, for each cluster, the CPU 21 selects a word to be displayed from the words included in the cluster (step S205). In step S205, from the words included in each cluster, n words or less are selected in the order of high occurrence frequency. Next, the CPU 21 determines the display size of the single character based on the frequency of occurrence of the single character for each single character selected in step S205 (step S206). In step S206, the higher the frequency of the single character, the larger the display size of the single character is determined.

Next, the CPU 21 generates screen data for displaying the results of the hierarchical cluster analysis (step S207). The screen data generated in step S207 includes m groups (represented by cloud shapes) having the size determined in step S204. Within each group, there are n or less words having the size determined in step S206. The single word is displayed in the group within the screen. After executing step S207, the CPU 21 ends the screen data generation processing.

Next, the CPU 21 causes the display unit 25 to display a screen based on the screen data generated in step S111 (step S112). Next, the CPU 21 receives an instruction from the user (step S113). Next, the CPU 21 proceeds to any one of steps S115 to S120 according to the type of the instruction received in step S113 (step S114).

When the instruction received in step S113 is "setting of the number of groups", the CPU 21 proceeds to step S115. In this case, the CPU 21 sets the group number m to a value instructed by the user (step S115), and proceeds to step S111. Thereafter, screen data is generated based on the set number of groups m, and a new screen is displayed. Thereby, the analysis result screen of the group including the designated number is displayed.

When the instruction received in step S113 is "setting of the maximum number of data in the group", the CPU 21 proceeds to step S116. In this case, the CPU 21 sets the maximum number of data n in the group to a value instructed by the user (step S116), and proceeds to step S111. Thereafter, screen data is generated according to the maximum number of data n in the set group, and a new screen is displayed. As a result, the analysis result screen in which the number of single words contained in each group is limited to a value specified below is displayed.

When the instruction received in step S113 is "setting of the analysis target period", the CPU 21 proceeds to step S117. In this case, the CPU 21 sets the analysis target period to the period instructed by the user (step S117), and proceeds to step S109. Thereafter, the hierarchical cluster analysis is performed with reference to the set analysis target period, and screen data for displaying a new analysis result is generated and a new screen is displayed. Thereby, the result of performing hierarchical cluster analysis on the words included in the text data within the specified analysis target period is displayed on the screen.

FIG. 11A is a diagram showing an analysis result screen before an analysis target period is set. FIG. 11B is a diagram showing an analysis result screen after the analysis target period is set. to The analysis result screen 61 before the setting shown in FIG. 11A displays the text data stores 5 of the input text data from 00:00 on January 1, 2014 to 24:00 on December 31, 2015. The results of hierarchical cluster analysis of the included words. In the analysis result screen 62 shown in FIG. 11B after the setting, the text data of the input text data 5 is displayed from 00:00 on March 1, 2014 to 24:00 on September 30, 2014. The results of hierarchical cluster analysis of the included words. The display content of the analysis result screen 61 is different from the display content of the analysis result screen 62. The user can easily recognize the temporal change of the hierarchical cluster analysis result by observing the analysis result screens before and after the analysis target period is set.

When the instruction received in step S113 is "except for a single word", the CPU 21 proceeds to step S118. In this case, the CPU 21 adds the specified word to the excluded word list (step S118), and proceeds to step S109. After that, the designated words are excluded and hierarchical cluster analysis is performed to generate screen data for displaying new analysis results and display the new screen. As a result, the result of performing hierarchical cluster analysis excluding the designated word is displayed on the screen.

FIG. 12A is a diagram showing an analysis result screen before single word exclusion is performed. FIG. 12B is a diagram showing an analysis result screen after single words are excluded. The user operates the mouse 29 and, after selecting a word to be excluded, instructs the word to be excluded. In the analysis result screen 63 before the single word exclusion shown in FIG. 12A, select "shakai (Society)", and select "except single word" from the menu. Thereafter, the result of the hierarchical cluster analysis except "shakai" is displayed on the screen. The analysis result screen 64 after excluding the single character shown in FIG. 12B displays “shingaku (promoting)” instead of “shakai”. Among the words included in the same cluster as "shakai", "shingaku" is the second word with the highest frequency after the 5 words displayed on the analysis result screen 63.

When the instruction received in step S113 is "synonym registration", the CPU 21 proceeds to step S119. In this case, the CPU 21 adds the indicated word to the currently used synonyms list (step S119), and proceeds to step S109. Thereafter, a hierarchical cluster analysis is performed in consideration of the indicated synonyms, and screen data for displaying a new analysis result is generated and a new screen is displayed. Thereby, the result of the hierarchical cluster analysis using the indicated word as a synonym is displayed on the screen.

FIG. 13A is a diagram showing an analysis result screen before synonyms are registered. FIG. 13B is a diagram showing an analysis result screen after synonyms are registered. The user operates the mouse 29 and, after selecting a plurality of words to be registered as synonyms, instructs the registration of synonyms. In the analysis result screen 65 before the synonyms registration shown in FIG. 13A, "daigakusei (university student)" and "gakusei (student)" are selected, and "synonym registration" is selected from the menu. After that, the results of hierarchical cluster analysis using "daigakusei" and "gakusei" as synonyms are displayed on the screen. In the analysis result screen 66 after the registration of the synonyms shown in FIG. 13B, "daigakusei" is displayed in a larger size than the analysis result screen 65, and "shingaku" is displayed instead of "gakusei". Based on the sum of the occurrence frequency of "daigakusei" and the occurrence frequency of "gakusei", "daigakusei" is displayed in a larger size than "daigakusei" in the analysis result screen 65.

When the instruction received in step S113 is "composite word registration", the CPU 21 proceeds to step S120. In this case, the CPU 21 adds the indicated word to the compound word list in use (step S120), and proceeds to step S109. After that, a hierarchical cluster analysis is performed in consideration of the indicated compound word, and screen data for displaying a new analysis result is generated, and a new screen is displayed. Thereby, the result of the hierarchical cluster analysis using the designated word as a compound word is displayed on the screen.

FIG. 14A is a diagram showing an analysis result screen before compound word registration is performed. FIG. 14B is a diagram showing an analysis result screen after compound word registration is performed. After the user operates the mouse 29 to select a plurality of words to be registered as a compound word, he instructs to perform "synonym registration". In the analysis result screen 67 before the compound word registration shown in FIG. 14A, "nintai" and "tsuyoi" are selected, and "composite word registration" is selected from a menu. Thereafter, hierarchical cluster analysis using "nintai" and "tsuyoi" as compound words is displayed on the screen. In the analysis result screen 68 after the compound word registration shown in FIG. 14B, "nintai" and "tsuyoi" are replaced, and "nintaizuyoi (high endurance)" is displayed in a size below "nintai" and "tsuyoi".

As shown above, the text exploration method of this embodiment includes: a text analysis step, which performs hierarchical cluster analysis on the words extracted from the input text data; a screen generation step, which is based on the analysis result of the text analysis step To generate screen data; and an analysis result display step that displays a screen based on the screen data. Screen generation step: According to the number of groups m and the maximum number of data n in the group, m clusters are obtained from the analysis results, and a group of words containing the words included in the clusters of n or less is displayed on the screen. Screen data. According to the text exploration method of the present embodiment, a group of words included in the text can be displayed on the screen according to a hierarchical cluster analysis result of the words included in the text data. The number of words included in the group is limited to n or less. Therefore, when the user sees the screen, he can intuitively understand the results of the hierarchical cluster analysis.

In addition, the words included in the group are selected from the words included in the cluster corresponding to the group in the order of high occurrence frequency. Therefore, within the group, words with high frequency among the words included in the cluster are displayed. Therefore, the user can easily To recognize the frequently occurring words contained in each cluster. In addition, the group has a total size corresponding to the appearance frequency of the words included in the cluster corresponding to the group in the screen. Therefore, the user can easily recognize clusters with a large total of the frequency of single words. In addition, the words included in the group have a size corresponding to the appearance frequency of the words in the screen. Therefore, the user can easily recognize words that appear frequently.

The text exploration method includes an instruction input step for inputting instructions from a user, and either the text analysis step or the screen generation step is executed based on the instructions input in the instruction input step. Therefore, according to the instructions from the user, the display mode of the results of the hierarchical cluster analysis can be switched. In particular, the instruction input step receives a setting instruction of the number of groups m, and the screen generation step generates screen data based on the number of groups m specified in the instruction input step. Thereby, according to an instruction from the user, the number of areas (the number of clusters) displayed on the screen is switched. In addition, the instruction input step receives the maximum number of data n in the group, and the screen generation step generates screen data according to the maximum number of data n in the group specified by the instruction input step. Thereby, according to the instruction from the user, the number of single words displayed in the area is switched.

In addition, the instruction input step receives instructions from the analysis target period, and the text analysis step performs hierarchical cluster analysis on words included in the text data in the text data during the analysis target period specified by the instruction input step. Therefore, the result of hierarchical cluster analysis on the words included in the text data within the analysis target period indicated by the user is displayed on the screen. Therefore, the user can easily recognize the temporal change of the results of the hierarchical cluster analysis. In addition, the instruction input step receives a setting instruction of the analysis purpose, and the text analysis step extracts a word corresponding to the type of the analysis purpose set in the instruction input step from the text data 5 to perform hierarchical cluster analysis. In this way, the type of words to be analyzed can be switched according to the analysis purpose indicated by the user, and The results of the hierarchical cluster analysis are displayed on the screen.

In addition, the instruction input step receives a single word exclusion instruction, and the character analysis step excludes the words indicated in the instruction input step, and performs hierarchical cluster analysis. In this way, the results of hierarchical cluster analysis can be displayed except for the words indicated by the user. In addition, the instruction input step receives a synonym registration instruction, and the character analysis step treats the plurality of words indicated in the instruction input step as the same word, and performs hierarchical cluster analysis. In this way, the multiple words indicated by the user can be regarded as the same word, and the result of the hierarchical cluster analysis can be displayed on the screen. In addition, the instruction input step receives a compound word registration instruction, and the character analysis step combines a plurality of words indicated in the instruction input step into one word, and performs hierarchical cluster analysis. In this way, a plurality of words instructed by the user can be combined into one word, and the result of the hierarchical cluster analysis can be displayed on the screen.

The screen generation step generates screen data, which is used to display an analysis result screen including a group and an analysis setting screen to set a display state of the analysis result screen. Therefore, the analysis result screen and the analysis setting screen are displayed. Therefore, the user can easily switch the display state of the result of performing the hierarchical cluster analysis using the analysis setting screen.

The character surveying program 31 of this embodiment and the character surveying device 10 of this embodiment have the same configuration as the character survey processing method of this embodiment, and exhibit the same effects.

According to the text exploration method, text exploration program, and text exploration device of this embodiment, the results of hierarchical cluster analysis of the words included in the text data can be used to make the group of words included in the cluster with the largest number of data or less Is displayed on the screen. Therefore, when users see the screen, they can intuitively understand the hierarchical cluster. The results of the analysis.

Furthermore, this case claims to file an application based on the priority of Japanese Patent Application No. 2016-145065 entitled "Text Exploration Method, Text Exploration Procedure, and Text Exploration Device" for the name of the invention filed on July 25, 2016. The contents of these applications are included in this application by reference.

40‧‧‧display

41‧‧‧ Analysis result screen

42‧‧‧ Analysis Setting Screen

Claims (25)

A text exploration method is to display the analysis result of text data on the screen; it is characterized by having: a text analysis step that performs hierarchical cluster analysis on the words extracted from the input text data; screen generation Step, which generates screen data according to the analysis result of the text analysis step; and analysis result display step, which displays the screen based on the screen data; the screen generation step is based on the number of groups and the maximum number of data in the group. The result of the analysis is to obtain a cluster of the above-mentioned number of groups, and to generate screen data for displaying on the screen a group containing the words included in the above-mentioned clusters below the maximum number of data. For example, the text exploration method of claim 1, wherein the words included in the above group are selected from the words included in the cluster corresponding to the above group in the order of high occurrence frequency. For example, the text exploration method of claim 2, wherein the group has a total size corresponding to the frequency of occurrence of the words included in the cluster corresponding to the group in the screen. For example, the text exploration method of claim 3, wherein the characters included in the group have a size corresponding to the frequency of occurrence of the characters in the screen. For example, the text exploration method of claim 1 further includes an instruction input step for inputting instructions from a user, and any one of the text analysis step and the screen generation step is based on the input in the instruction input step. Instructions are executed. For example, the text exploration method of item 5, wherein the instruction input step is received on the In the setting instruction for the number of groups, the screen generation step generates the screen data according to the number of groups set in the instruction input step. For example, the text exploration method of claim 5, wherein the instruction input step receives the setting instruction of the maximum number of data, and the screen generation step generates the screen data according to the maximum number of data set in the instruction input step. For example, the text exploration method of claim 5, wherein the instruction input step receives the setting instruction of the analysis target period, and the text analysis step includes the text data included in the text data during the analysis target period set in the instruction input step. Single word, perform the above-mentioned hierarchical cluster analysis. For example, the text exploration method of claim 5, wherein the instruction input step receives a setting instruction of the analysis purpose, and the text analysis step extracts from the text data a word corresponding to the type of analysis purpose set in the instruction input step to Perform the hierarchical cluster analysis described above. For example, the text exploration method of claim 5, wherein the instruction input step receives a single-word exclusion instruction, and the text analysis step excludes the word indicated by the instruction input step and performs the hierarchical cluster analysis. For example, the text exploration method of claim 5, wherein the instruction input step receives a synonym registration instruction, and the text analysis step will include a plurality of words indicated in the instruction input step. Consider the same word and perform the hierarchical cluster analysis described above. For example, the text exploration method of claim 5, wherein the instruction input step receives a compound word registration instruction, the text analysis step combines the plurality of words indicated in the instruction input step into one word, and performs the hierarchical cluster analysis. For example, the text exploration method of claim 1, wherein the screen generation step generates screen data, and the screen data is used to display the analysis result screen including the group and the analysis setting to set the display state of the analysis result screen. Screen. A computer-readable recording medium records a text exploration program that displays analysis results of text data on a screen. The computer-readable recording medium is characterized in that the CPU uses a memory to cause a computer to execute the following steps: a text analysis step, which is performed on a self-input The words extracted from the text data are subjected to hierarchical cluster analysis; the screen generation step generates screen data according to the analysis result of the text analysis step; and the analysis result display step displays the screen according to the screen data; the above screen The generating step obtains the cluster of the above-mentioned group number from the analysis result according to the number of the group and the maximum number of data in the group, and generates a group display for displaying the word including the single word included in the above-mentioned cluster below the maximum number of data Picture data on the screen. For example, the computer-readable recording medium of claim 14, wherein the words included in the above group are selected from the words included in the cluster corresponding to the above group in order of high frequency of occurrence. For example, the computer-readable recording medium of item 15, wherein the above group has an appearance corresponding to the word contained in the cluster corresponding to the above group in the above screen. The total size of the frequencies. For example, the computer-readable recording medium of claim 16, wherein the words included in the group have a size corresponding to the frequency of appearance of the words in the screen. For example, the computer-readable recording medium of claim 14, wherein the computer is further caused to execute the instruction input step for inputting instructions from the user, and any one of the text analysis step and the screen generation step is based on The instruction input in the instruction input step is executed. For example, the computer-readable recording medium of claim 14, wherein the screen generation step generates screen data, and the screen data is used to display the analysis result screen including the group and to set the display status of the analysis result screen. The same analysis setting screen. A text exploration device, which displays the analysis results of text data on a screen, is characterized in that it includes: a text analysis unit that performs hierarchical cluster analysis on words extracted from input text data; screen generation Unit, which generates screen data based on the analysis result of the text analysis unit; and analysis result display unit, which displays the screen based on the screen data; the screen generation unit automatically The above analysis results obtain the cluster of the number of groups, and generate screen data for displaying on the screen a group of words containing the single word included in the cluster below the maximum number of data. For example, the character exploration device of claim 20, wherein the words included in the group are selected from the words included in the cluster corresponding to the group in the order of high occurrence frequency. For example, the text exploration device of claim 21, wherein the group has a total size corresponding to the frequency of occurrence of the words included in the cluster corresponding to the group in the screen. For example, the character exploration device of claim 22, wherein the words included in the group have a size corresponding to the appearance frequency of the words in the screen. For example, the text surveying device of claim 20 further includes an instruction input unit for inputting instructions from a user, and any one of the text analysis unit and the screen generation unit, according to the instructions input in the instruction input unit. Come to action. For example, the text exploration device of claim 20, wherein the screen generation unit generates screen data, and the screen data is used to display an analysis result screen including the group and an analysis setting to set a display state of the analysis result screen Screen.