JP2003044485A - Knowledge analysis system and method for setting maximum number of top-level clusters in the system - Google Patents

Abstract

(57) [Summary] [Problem] To provide a knowledge analysis system capable of setting the maximum number of clusters in the highest hierarchy. Kind Code: A1 A knowledge server stores in a knowledge database as a separately indexed document a document requested to be registered in a document database from a client terminal.
It supports the understanding of the tendency to be read from the knowledge group accumulated in the H.225. Then, the knowledge server 122
Has a mechanism that allows the user to set the maximum number of clusters in the highest hierarchy in order to prevent an unlimited increase in top clusters with the increase in accumulated knowledge. Perform a so-called clustering process to sort the obtained knowledge group into knowledge clusters,
The result is stored in the analysis result database 1227.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a knowledge analysis system used in a knowledge management system and a maximum number setting method of clusters in the system,
In particular, the present invention relates to a knowledge analysis system capable of setting the maximum number of clusters in the highest hierarchy and a method of setting the maximum number of highest clusters in the system.

[0002]

2. Description of the Related Art In recent years, the introduction of groupware for sharing information among a plurality of users has been promoted mainly in companies. E-mail systems and workflow systems are known as typical groupware,
Recently, a knowledge management system to support sharing of knowledge and information has begun to be developed.

This knowledge management system is based on W
This is for accumulating and managing personal know-how as a knowledge database in addition to eb information and electronic file information. By combining with a search function such as natural language search, efficient use of knowledge and information is possible. Becomes

By the way, in such a knowledge management system, how to collect and accumulate knowledge such as individual know-how is important. This is because personal know-how and other knowledge is so-called tacit knowledge and is not formalized like Web information or electronic file information, so it is difficult to automatically collect and store it. .

Therefore, recently, the development of a knowledge management system having a knowledge accumulation support function is required. Utilizing the knowledge as tacit knowledge in the same way as formalized knowledge such as Web information and electronic file information by realizing a mechanism that automatically collects and accumulates knowledge such as individual know-how. Is possible.

In addition, a knowledge management system for easily retrieving the knowledge and information accumulated in this way is also being developed in parallel. A typical example is a natural language search system that has a knowledge search support function that searches for useful knowledge and information by inputting a natural language question sentence.

[0007]

By the way, in this kind of knowledge management system, it is possible to easily organize or browse knowledge or information, or to retrieve any kind of knowledge information for a user who uses it for the first time. It is strongly required to systematize the knowledge so as to effectively utilize the knowledge data, such as presenting the question. Along with this, the accumulated knowledge is categorized and hierarchically classified.

However, in the past, as the accumulated knowledge increased, the total number of categories also increased indefinitely, so that the scattering of knowledge became remarkable, and it became difficult to obtain useful knowledge analysis results. there were.

The present invention has been made in consideration of such circumstances, and a knowledge analysis system capable of setting the maximum number of clusters in the highest hierarchy and a method of setting the maximum number of highest clusters in the system The purpose is to provide.

[0010]

In order to achieve the above-mentioned object, the present invention is configured to be connectable to a plurality of client terminals via a network, and the client terminals corresponding to the knowledge accumulated in the knowledge database from each client terminal. A knowledge analysis system for supporting analysis, comprising: access control means for authenticating a user to an access request source client terminal to permit knowledge analysis from the client terminal; and access permission by the access control means. And a knowledge analysis means for creating a cluster database for each cluster defined for each category of knowledge accumulated in the knowledge database according to analysis conditions. The knowledge analysis means is the cluster database When creating comprises means for setting a maximum number of clusters formed on the highest layer, to provide knowledge analysis system being characterized in that to be able to limit the number of clusters highest layer.

This knowledge analysis system has a mechanism for setting the maximum number of clusters in the highest layer, so that the number of clusters in the highest layer can be limited, and as the accumulated knowledge increases, It prevents the top-level cluster from growing indefinitely, which prevents the sub-clusters from exploding. That is, the problem that knowledge is scattered and it becomes difficult to obtain a useful analysis result is solved, and a simple and effective knowledge analysis is realized.

[0012]

DETAILED DESCRIPTION OF THE INVENTION An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows the configuration of a knowledge analysis system according to the embodiment of the present invention. This knowledge analysis system is used as a knowledge management system having a knowledge analysis function, and a knowledge database 1225 that can be commonly used by a plurality of client terminals 11 is provided.
Use to analyze knowledge. Before describing a specific configuration, first, an outline of the knowledge analysis system according to this embodiment will be described with reference to FIGS. 2 to 9.

The knowledge analysis system of this embodiment is a system for assisting in grasping a tendency or the like that can be read from a miscellaneous document group collected for a certain purpose, that is, a knowledge group. This knowledge analysis system is composed of the following three functions.

(1) Automatic creation of knowledge cluster Knowledge clusters (groups of knowledge) are collected by collecting similar knowledge.
A function for automatically sorting and editing. (2) Display of knowledge cluster list Function to compare the number of knowledge among the classified knowledge clusters. (3) Time-series graph display of knowledge in a knowledge cluster A function of viewing the tendency of knowledge generation in a certain knowledge cluster. With these, ・ Analysis of questionnaires ・ Trend analysis of questions to the help desk ・ Classification of ideas ・ Selling products from sales daily reports and factor analysis Extracting the latest topics from the bulletin board Unexpected discovery / new awareness (new knowledge It can be used for applications such as creation.

The respective functions will be described below.

<Automatic Creation of Document Cluster> FIG. 2 shows the concept of automatic creation of a knowledge cluster, and FIG. 3 shows the flow of clustering. The knowledge cluster (knowledge group) is created by classifying a large amount of knowledge collected in the database of the knowledge management system into groups with similar contents based on whether similar words are used.
Further, it is possible to classify the clusters into clusters in which there is a cluster, and the created knowledge cluster group can be saved with a name (classification name) as "classification".

(1) Designation when searching There are the following two specifications at the time of search.

Search condition: keyword search. -The number of knowledge items to be clustered (how many higher-ranked search scores are used). (2) Designated item A There are the following nine designated items A.

Designation of important words / unnecessary words / synonyms. -Important words: words that are assumed to be important for analysis. Has priority in clustering. -Unnecessary words: words unnecessary for analysis. Ignored by clustering. -Synonyms: Words that have the same meaning in analysis. Treated as the same word in clustering.

The number of knowledge clusters to be created (first hierarchy).・ Whether the number of layers should be one or the number of layers should not be specified.・ Knowledge can only fit in one cluster (1: 1),
Will there be more than one (1 to n)?・ Whether the labels should be only nouns or other parts of speech. -Name of the entire cluster (classification name). (3) Designation item B (designation at the time of reclustering). The designated item B has the following four items.

Designation of fixed clusters (clusters that are not broken even during re-clustering). -Designation of clusters to be integrated (two or more clusters → one cluster).・ Designation of important / unnecessary words. -Number of knowledge clusters to be created (first layer: number of clusters formed after re-clustering).

A large amount of knowledge is stored in the knowledge database 1225 of FIG. 2 without any ordering or organization. When the knowledge condition to be clustered is specified in the knowledge database 1225 and searched (step A1 in FIG. 3), the set of searched knowledge is determined as shown in FIG. 2 (step A2). When a condition for clustering is designated (step A3), clustering (automatic grouping) is executed to generate a cluster name, determine a hierarchical structure, and allocate knowledge (step A4), so that knowledge cluster 1 to knowledge cluster n in FIG. It is created and saved as the first knowledge cluster group (step A5, step A6).

Although the cluster group is first created in this way, if the end user wants to perform clustering again under different conditions (YES in step A7), the specified conditions are adjusted (the specified item B) as necessary (step A).
8) Perform re-clustering (this is called feedback) (step A9). As a result, a knowledge cluster group adjusted as a result of reclustering is created and saved (step A10, step A6). On the other hand, when reclustering is not necessary from the beginning or when further reclustering is unnecessary. (NO in step A7), the clustering process ends.

<Display of Knowledge Cluster List> (1) Hierarchy non-display & graph non-display is performed in the order of the number of knowledge items. FIG. 4 shows an example of a display screen of a list of all clusters in order of the number of cases. Displaying a list of all clusters ignoring the hierarchy and the specific hierarchy (first hierarchy, second hierarchy)
There are two types of list display in the hierarchy, ..., The nth hierarchy. The screen of FIG. 4 will be described below.

Cluster name: The cluster name is displayed. -Number of cases: The number of cases of knowledge belonging to the cluster is displayed. -Keyword: The keyword of the cluster is displayed. (2) Hierarchy non-display & graph display FIG. 5 is a screen in which all clusters are displayed in a list regardless of the hierarchy. Displaying a list of all clusters ignoring the hierarchy,
There are two types: list display in a specific hierarchy. Below, FIG.
Screen is explained.

Cluster name: The cluster name is displayed. -Number of cases: The number of cases of knowledge belonging to the cluster is displayed. -Keyword: The keyword of the cluster is displayed. -Graph: The number of knowledge items is displayed in a bar graph.

(3) Hierarchical display & non-graph display There are the following two methods of displaying the hierarchical relationship for all clusters.

(A) Display of entire hierarchy (b) Display with fixed hierarchy and display without hierarchy FIG. 6 shows a screen display example of the entire hierarchy. this is,
It is a screen in which the analysis result is displayed only in a hierarchical structure. The screen will be described below.

Cluster: The analyzed cluster and the keywords of the cluster are displayed in a hierarchical structure.

Number of cases: The number of cases of knowledge under the cluster is displayed.

(4) Hierarchical display & graph display There are the following two methods for displaying the hierarchical relationship for all clusters.

(A) Display of entire hierarchy (b) Display with fixed hierarchy and display without hierarchy FIG. 7 shows a screen display example of the entire hierarchy. this is,
It is a screen in which the analysis result is displayed in a hierarchical structure and a graph.
The screen will be described below.

Cluster: The analyzed cluster and the keywords of the cluster are displayed in a hierarchical structure. -Number of cases: The number of knowledge cases under the cluster is displayed.

Graph: The number of knowledge cases is displayed as a bar graph. Also, each layer is displayed in a different color. <Time-series graph display of knowledge in knowledge cluster> The number of knowledge items registered in the knowledge cluster is displayed in a time-series graph (daily or monthly).

(1) Time-series analysis display This is a function of displaying the registered knowledge under the cluster in units of months or days according to the registered time information. When the screen starts,
The graph is not displayed, but the graph of the specified range is displayed by specifying the display unit and display range. Figure 8
Is an example displayed on a monthly basis. The screen will be described below.

Display unit: The unit for displaying the graph is designated as either a monthly unit or a daily unit. The default is "monthly". -Display range: Specify the display range. -Graph display: In the case of monthly units, the number of knowledge items that match the specified year of the display range is aggregated on a monthly basis and displayed as a graph.
At this time, even if a numerical value is entered in the "month" item of the display range, it is not reflected in the graph display. On the other hand, in the case of daily units, the number of knowledge items that match the specified year and month of the display range is aggregated on a daily basis and displayed as a graph.

(2) Classification list The result of clustering (a group of knowledge clusters) is stored as a classification. A screen for displaying a list of saved classifications is a classification list. The screen of FIG. 9 is the initial screen of the knowledge analysis function.

By selecting a classification from the screen shown in FIG. 9, a list of clusters in that classification can be viewed. New creation can also be done from this screen. When the knowledge analysis function is activated, the screen shown in FIG. 9 is displayed and the analysis result is selected. The screen will be described below.

Analysis result: The name of the analysis result is displayed. This name is specified by the user when creating the analysis result. -Knowledge DB: The name of the knowledge DB used for analysis is displayed. -Number of cases: The total number of cases of the analyzed DB is displayed. -Highest cluster number: The number of clusters in the first layer is displayed. -Update date: The date when the analysis result was last updated is displayed. -Created date: The date when the analysis result was created is displayed. <System Configuration> Next, the system configuration of the knowledge analysis system according to this embodiment will be described with reference to FIG.

The knowledge analysis system of this embodiment is realized by a plurality of client terminals 11 and a server computer 12 which can be connected to each other via a computer network 13 such as a LAN. Server computer 12
Although not shown, the client terminal 11 and the client terminal 11 are each provided with an input / output device having a CPU, a main memory, a magnetic disk device as a storage device, an input unit such as a keyboard and a mouse, and a display unit such as a display. There is.

On the client terminal 11, the Web browser 111 is operating. URL indicating the resource for knowledge analysis constructed on the server computer 12 (Un
By specifying the iform Resource Locator) from the Web browser 111, the knowledge analysis process can be used from each client terminal 11.

The knowledge analysis function of the server computer 12 is mainly performed by the control module 121 of the Web server 121.
1, the registration module 1221 of the knowledge server 122,
It is realized by software such as the search module 1222 and the clustering module 1223, and management information and actual data used by these software for knowledge analysis. The management information includes login management information 1212 for performing user authentication for each client terminal 11. As actual data, there are a document database 1224, a knowledge database 1225, an intermediate processing database 1226, and an analysis result database 1227.

The control module 1211 is for controlling the entire operation relating to knowledge analysis, and is the core program of the knowledge analysis system, the knowledge server 1.
22 and the web server 121, and the W
It has a user authentication function when each client terminal 11 logs in to the knowledge server 122 through the eb server 121. For this user authentication, the control module 12
Reference numeral 11 manages login management information 1212. The login management information 1212 stores the user ID and password of each user participating in this knowledge analysis system. By this user authentication, control of permission / prohibition of access to the knowledge server 122 for knowledge analysis from each client terminal 11 is performed.

The knowledge server 122 manages and operates the knowledge database 1225 and the analysis result database 1227, which are the knowledge that can be analyzed by a plurality of client terminals 11, and are based on the conditions specified by each client terminal 11. The knowledge analysis results are classified and accumulated as knowledge clusters.

The maximum number cluster setting process in the knowledge analysis system executed by the clustering module 1223 of the knowledge server 122 will be described below. The process of setting the maximum number of top-level clusters is to prevent the number of top-level clusters from increasing indefinitely as the accumulated knowledge increases. This is a function to set the maximum number of clusters in the upper layer.

First, FIG. 10 and FIG. 11 will be described.
FIG. 10 is a knowledge database selection screen. The following items are displayed in FIG.

Knowledge DB (comment / number): name, comment and number of knowledge database 1225. The knowledge database 1225 is provided as a pair with the document database 1224, and stores documents stored in the document database 1224, that is, an index of words and phrases included in knowledge. Client terminal 1
When the storage module 1221 of the knowledge server 122 stores this document in the specified document database 1224, the registration module 1221 of the knowledge server 122 stores the document in the specified document database 1224. For example, morphological analysis is performed to create an index of words included in the document. Then, the created index is stored in the knowledge database 1225.

FIG. 11 is a screen for designating conditions for analyzing the knowledge database 1225 selected in FIG. The following items are displayed in FIG.

Analysis target DB: The names, comments and total number of all knowledge data selected on the knowledge database selection screen of FIG. 10 are displayed. -Analysis result name: Enter the name for saving the analysis result. If the real name overlaps with the already saved analysis result, a warning message is displayed when <Execution> is selected, and if <Save> in the message is selected, the existing analysis result is overwritten and saved.・ Analysis period: Enter the period to be analyzed. If there is no input, all knowledge will be analyzed. -Refinement keyword: Enter a keyword that is a keyword for analysis. If there is no input, all knowledge will be analyzed. -Number of narrowed down cases: Enter the maximum number of cases registered in the analysis result. When the analysis result is larger than the number of analysis cases, the knowledge for the specified number of cases is extracted from the one having the highest conformity with the analysis conditions. If there is no input, all knowledge that meets the analysis conditions is extracted. -Number of layers: Specify the number of layers for clustering. The default is "multi-level". -Knowledge duplication: Specify whether or not to allow one knowledge to be registered in multiple clusters in duplicate. If "Yes" is selected, each knowledge is registered in all related clusters, and as a result, the same knowledge exists in different clusters. Therefore, the total of the analysis results may exceed the sum of all the numbers in the analysis source DB. On the other hand, when “none” is selected, each knowledge is registered in only one cluster having the strongest association. The default is "Yes". -Maximum number of top level clusters: Specify the maximum number of clusters created in the top level hierarchy. If there is no input,
Clustering is performed with no designation.

Key word: Input a key word for analysis. When entering multiple items, separate them with spaces.・ Unnecessary words: Input unnecessary words for analysis. When entering multiple items, separate them with spaces.・ Synonyms: Enter synonyms for analysis. Describe synonyms by connecting them with “=” and separate them with a semicolon (;). It is also possible to define three or more synonyms for one item. Further, the defined synonyms are aggregated into the word described on the leftmost side in the analysis result.

FIG. 12 is a flow chart showing a procedure for executing knowledge analysis on the knowledge database 1225 selected in FIG. 10 based on the maximum number of highest clusters designated on the analysis condition designation screen of FIG.

When the user makes a login request to the control module 1211 of the server computer 12 using the Web browser 111 (step B1), the control module 121
1 accesses the login management information 1212 to check whether or not the user ID and password input by the user are registered (step B2), and performs user authentication for determining whether or not to permit this login. Perform (step B03). If the user ID and password are not registered in the login management information 1212 and the login fails (NO in step B3), the control module 1211 notifies the Web server 121 of the login failure.
The process is returned to the Web browser 111 through and the process ends (step B4).

On the other hand, if the user ID and password are registered in the login management information 1212 and the login is successful (YES in step B3), the clustering module 1223 of the knowledge server 122 selects the knowledge database shown in FIG. A screen is displayed (step B5). When this knowledge database selection screen is displayed, the user selects the analysis target knowledge database 1225 on the screen (step B6). When this selection is made, the clustering module 1223 of the knowledge server 122 this time displays the analysis condition designation screen shown in FIG. 11 (step B7).

When this analysis condition designation screen is displayed, the user sets the analysis conditions on this screen (step B8). When this setting is made, Knowledge Server 1
22 of the search module 1222, for the knowledge database 1225 selected on the knowledge database selection screen,
The intermediate processing database 12 is searched by using the important words, unnecessary words, and synonyms specified on the analysis condition specification screen.
26 is created (step B9). Next, the clustering module 1223 of the knowledge server 122 checks whether or not the maximum number of the highest-level clusters is designated on the analysis condition setting screen (step B10), and if it is designated (YES in step B10), the highest number is determined. The clustering process is executed while considering the maximum number of upper clusters (step B11). On the other hand, if not specified (NO in step B10), the clustering process is executed without limiting the maximum number of highest clusters (step B12).

As described above, since this knowledge analysis system has a mechanism for setting the maximum number of clusters in the highest hierarchy, it is possible to increase the number of highest clusters without limit as the accumulated knowledge increases. Prevent and by this
Prevent an explosion of sub-clusters. That is, the problem that knowledge is scattered and it becomes difficult to obtain a useful analysis result is solved, and a simple and effective knowledge analysis is realized.

The invention of the present application is not limited to the above-described embodiment, and can be variously modified at the stage of implementation without departing from the scope of the invention. Furthermore, the embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some constituent elements are deleted from all the constituent elements shown in the embodiment, the problem described in the section of the problem to be solved by the invention can be solved,
When the effects described in the section of the effects of the invention can be obtained, a structure in which this constituent element is deleted can be extracted as the invention.

[0058]

As described above in detail, according to the present invention, it is possible to limit the number of clusters in the highest hierarchy by having a mechanism for setting the maximum number of clusters in the highest hierarchy. It prevents the top-level cluster from growing indefinitely as the accumulated knowledge increases.
Prevent an explosion of sub-clusters. That is, the problem that knowledge is scattered and it becomes difficult to obtain a useful analysis result is solved, and a simple and effective knowledge analysis is realized.

[Brief description of drawings]

FIG. 1 is a block diagram showing a system configuration of a knowledge analysis system according to an embodiment of the present invention.

FIG. 2 is a view for explaining the concept of automatically creating a knowledge cluster in the knowledge analysis system of the same embodiment.

FIG. 3 is a follow chart showing a flow of clustering in the knowledge analysis system of the embodiment.

FIG. 4 is an exemplary view showing an example of a cluster list display screen (hierarchy non-display & graph non-display) in the knowledge analysis system of the embodiment.

FIG. 5 is an exemplary view showing an example of a cluster list display screen (hierarchical non-display & graph display) in the knowledge analysis system of the embodiment.

FIG. 6 is a display screen (hierarchical display & graph non-display) of the entire cluster hierarchical relationship in the knowledge analysis system of the same embodiment.
The figure which shows an example.

FIG. 7 is an exemplary view showing an example of a display screen (hierarchical display & graph display) of the entire cluster hierarchical relationship in the knowledge analysis system of the embodiment.

FIG. 8 is a diagram showing an example of a time-series display screen of each cluster in the knowledge analysis system of the same embodiment.

FIG. 9 is an exemplary view showing an example of a classification list display screen in the knowledge analysis system of the embodiment.

FIG. 10 is an exemplary view showing an example of a knowledge database selection screen in the knowledge analysis system of the embodiment.

FIG. 11 is an exemplary view showing an example of an analysis condition designation screen in the knowledge analysis system of the embodiment.

FIG. 12 shows a procedure for executing knowledge analysis based on the maximum number of highest clusters designated on the analysis condition designation screen for the knowledge database selected on the knowledge database selection screen in the knowledge analysis system of the embodiment. flowchart.

[Explanation of symbols]

11 ... Client terminal 12 ... Server computer 13 ... LAN 111 ... Web browser 121 ... Web server 122 ... Knowledge Server 1211 ... Control module 1212 ... Login management information 1221 ... Registration module 1222 ... Search module 1223 ... Clustering module 1224 ... Document database 1225 ... Knowledge database 1226 ... Database for intermediate processing 1227 ... Analysis result database

Continued front page    (72) Inventor Katsuhiko Kochio             2-9 Suehiro-cho, Ome City, Tokyo Stock Market             Company Toshiba Ome Factory (72) Inventor Kazuhiko Atsumi             2-9 Suehiro-cho, Ome City, Tokyo Stock Market             Company Toshiba Ome Factory (72) Inventor Junya Sasaki             2-9 Suehiro-cho, Ome City, Tokyo Stock Market             Company Toshiba Ome Factory (72) Inventor Kazunori Shimakawa             2-9 Suehiro-cho, Ome City, Tokyo Stock Market             Company Toshiba Ome Factory F-term (reference) 5B017 AA07 BA05 CA16                 5B075 KK02 KK42 ND03 NR12                 5B082 AA09 EA12

Claims (5)

[Claims] 1. A knowledge analysis system configured to be connectable to a plurality of client terminals via a network and supporting analysis from each client terminal with respect to knowledge accumulated in a knowledge database, wherein the knowledge from the client terminals is used. An access control means for authenticating a user to an access request source client terminal to permit analysis; and a cluster database used for knowledge analysis from each client terminal permitted to access by the access control means, And a knowledge analysis means for creating a cluster database for each cluster defined for each category according to the analysis condition based on the knowledge accumulated in the knowledge database, wherein the knowledge analysis means creates a cluster database at the highest level when the cluster database is created. Maximum number of clusters formed And means for setting,
A knowledge analysis system characterized in that the number of clusters in the highest hierarchy can be limited. 2. The knowledge analysis means according to claim 1, further comprising means for displaying a screen for designating the analysis condition, and setting the maximum number of the clusters from the screen. Analysis system. 3. A method for setting the maximum number of highest clusters in a knowledge analysis system configured to be connectable to a plurality of client terminals via a network and supporting analysis from each client terminal with respect to knowledge accumulated in a knowledge database. An access control step of performing user authentication to the client terminal of the access request source in order to permit the knowledge analysis from the client terminal; and a knowledge analysis from each client terminal permitted to access by the access control step. And a knowledge analysis step of creating a cluster database for each cluster defined for each category of knowledge accumulated in the knowledge database according to analysis conditions, the knowledge analysis step comprising: Base Maximum number setting method when the creation of top level clusters in the knowledge analysis system characterized by having a maximum number setting step of setting a maximum number of clusters formed on the highest layer. 4. The method for setting the maximum number of uppermost clusters in the knowledge analysis system according to claim 3, wherein the knowledge analysis step sets the analysis condition and the maximum number of clusters from a display screen. 5. A server computer configured to be connectable to a plurality of client terminals via a network,
A program for realizing the maximum number setting function of the highest cluster in the knowledge analysis system that supports the analysis from each client terminal for the knowledge accumulated in the knowledge database, wherein the server computer stores the knowledge from the client terminal. An access control means for authenticating a user to an access request source client terminal to permit analysis; and a cluster database used for knowledge analysis from each client terminal permitted to access by the access control means, The knowledge accumulated in the knowledge database is made to function as knowledge analysis means for creating a cluster database for each cluster defined for each category according to the analysis condition, and the knowledge analysis means forms the highest hierarchy when the cluster database is created. It A program characterized by executing the maximum number setting procedure to set the maximum number of clusters that.