JP2000172696A - Document managing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To extract the group of series of operation histories performed, based on the intention of a user out of the huge quantity of operation histories generated and stored at random with few processings. SOLUTION: An operation monitor part 5 monitors user operation inputted from an input device 1, extracts the preparing/updating/reference processing of a document or document output processing, such as transmission/printing and records the document name of an operation object, time when the operation occurs, the identifier of a user who performs the operation and the user identifier at the transmission destination or the like in an operation history preserving part 6 as the operation history. Then, a history cluster preparing part 7 analyzes the operation histories stored in the operation history preserving part 6, for example divides it into small history sequences for each user time sequentially arranged within a predetermined timewise range, and records this divided operation history group in a cluster preserving part 8 as a cluster.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document management system having a function of accumulating a history of operations on a document shared by a plurality of users, and more particularly, to a huge amount of data generated and accumulated at random. The present invention relates to a document management system capable of extracting a series of operation histories (business sequences) performed based on a user's intention from an operation history with a small processing amount.

[0002]

2. Description of the Related Art With the spread of large-capacity storage devices and networks, the use of digitized documents has become commonplace in modern offices. The ability to effectively retrieve electronic documents and how effectively they can use electronic documents has a significant effect on business efficiency.

The conventional file management method using a tree structure has insufficient functions for managing documents in office work. For example, a search function is supplemented in order to supplement the information management function and improve the work efficiency. A high-performance full-text search tool for documents using natural language and synonymous search, and a workflow system for improving work efficiency using predefined document flows, that is, adding a flow management function Many document management application programs have been developed.

[0004] These application programs are:
Although it has been effective to some extent as a tool to help search and reuse documents, it does not fully meet the needs of users and needs features that cannot be covered by these application programs. For example, a full-text search application program for a document cannot perform a search if the user does not think of a keyword included in the document.
Further, in the workflow system, the method of using the document must be clearly defined in advance, and it is difficult to use the document in daily office work in which atypical documents are often used.

[0005] Recently, attempts have been made to support the use of documents by using a user's document operation history. This attempt was based on a "document usage history" that was not available with traditional document management methods.
To search for a document using the information described in, for example, Japanese Patent Laid-Open No. 6-342.
In the document management apparatus described in Japanese Patent No. 451, by using the recorded user's document operation history to search for history information such as when a user referred to the document and when, the user can not come up with a search keyword. Even allows you to find the documentation you need.

[0006]

In the above-described method, a user who searches for a document searches for an operation history by guessing who and how the document was used. From the user's perspective, the usage histories such as "In what order did A access the document?" And "What information did you refer to when creating the document?" Since it is easier to use information such as the intention of operation at the time of operation, it is easy to use it.In addition to simply accumulating operation history, the user who performed the operation can summarize his own operation history according to purpose, A system has also been proposed that allows to leave a comment on. However, the method of having the user leave a comment has a problem that the burden on the user is heavy and the description of the comment frequently occurs.

On the other hand, as a method of analyzing the history, a method of applying a data mining method to find a frequently occurring sequential pattern from an event log,
Similarly, there are methods that use data mining techniques to find correlations between event logs.However, these methods use a huge number of combinations of event logs, Because it is used as a means of analyzing historical data, whether it occurs at a high frequency, the load for pattern detection is large, discontinuous sequences cannot be found, and infrequent correlations are difficult to find. However, it has been difficult to apply this method to a document management system that requires analysis of history data in which operation histories performed by a plurality of users with a plurality of intentions are mixed.

The present invention has been made in view of such circumstances, and provides a document management system capable of extracting a dependency of an operation history on a document with a small processing amount without burdening a user. With the goal.

[0009]

In order to achieve the above object, according to the present invention, when a history of operations on a document is viewed for each user, an operation within a short time range is associated with a specific operation. Noting that the operation is limited to the operations performed, continuous history data is divided into continuous operation sequences (clusters) for each user in a short time range.
The similarity between the created operation sequences (clusters) is determined, and the operation sequences having a high similarity (considered to have been performed for the same purpose) are connected to each other, and the user operation in a longer time range is performed. A history column (case) is created.

According to the present invention, without burden on the user,
In addition, to extract a series of operation histories (clusters and cases serving as business sequences) performed based on a user's intention from an enormous amount of operation histories with a processing amount much smaller than data mining. Becomes possible.

[0011]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a functional block diagram of the information system according to this embodiment.

The input device 1 is a keyboard, a mouse or an operation terminal connected via a network for inputting a user's operation to the system. Display device 2
Is a display device or the like that displays a result of an operation or a document to a user who has performed an operation based on a user operation. The output device 3 is a printer, a facsimile or a network device for outputting a document to the outside of the system based on a user operation.

The document management unit 4 is a normal operating system for managing storage areas and controlling input / output devices and display devices. The document management unit 4 is connected to a plurality of input / output devices via a network and is shared by a plurality of users. Has been used. The document management unit 4 includes the input device 1
Based on the user's operation, the user performs creation / edit / reference of the document and display / output of the operation result to the display device 2 or the output device 3. The document management unit 4 manages attributes such as file names and directory information of documents and information such as user groups. Further, the document management unit 4 is extended so that a user operation on a document can be monitored.

The operation monitor unit 5 monitors a user operation input from the input device 1 and performs a document creation / update / reference process, a document output process such as sending a document by e-mail / fax or printing a document. The extracted document name, the time at which the operation occurred, the identifier of the user who performed the operation, the user identifier of the transmission destination, and the like are recorded as the operation history in the operation history storage unit 6 described later.

The operation history storage unit 6 records the operation history detected by the operation monitor unit 5. FIG. 2 shows an example of recorded operation history data. In FIG. 2, a history ID (a1) is a number assigned to manage the operation history. The operation type (a2) represents the type of the recorded operation history, and includes “creation”, “update”, “reference”, “delete”,
The items of the operation performed by the user such as “print” or “send mail” are stored in the order in which they occur.

The document name (a3) is the name of the document to be operated recorded together with the operation history. In this example, the directory and the file name in the file system of the operating system are recorded. The operation date and time (a4) records the time at which the operation occurred.
The user name (a5) is the user identifier of the user who performed the operation left in the operation history, and is obtained from the user information managed by the document management unit 4.

The history cluster creating unit 7 analyzes and divides the operation histories stored in the operation history storage unit 6 and converts them into clusters, which are small history strings arranged in time order. The cluster is intended to summarize the history of processing performed by the user for the same purpose. According to the present invention, information as to why the user performed the operation is not left in the history data. However, even if the user performs a plurality of tasks in parallel, focusing on the operation in a short time range, Focusing on the characteristics of document operations, in which operations for specific purposes are often performed in a concentrated manner, the operation history sequence in which the operation contents of a plurality of users are recorded is divided by the user who performed the operation, and furthermore, A process for dividing the operation history into history strings generated within a short time interval is performed.

If the history is divided merely by focusing on the time range, the history of operations performed with different intents is likely to be included in the same cluster (for example, “another mail is generated while creating a report”). Received, the document creation business was temporarily suspended and a reply mail was created. "), The history cluster creating unit 7 empirically performs the user operation. Detects specific operation histories, such as document save operations, e-mail send operations, and document print operations that are often performed when changing the purpose of a document, and divides the operation history even when these operations occur Is performed.

The cluster created by the history cluster creating unit 7 is recorded in the cluster saving unit 8. FIG. 3 shows an example of the created cluster. In FIG. 3, a cluster ID (b1) is a number assigned to manage a group of operation histories grouped as a cluster. For example, the cluster ID 121 is a collection of operations performed by Mr. Suzuki from 10:30 to 10:43. Even if the operation is temporally close, the operation performed by another user is recorded as another cluster (cluster I
D122, 123). Although the operation of Mr. Suzuki included in the cluster ID 124 is close to the time of occurrence of the operation included in the cluster ID 121, the operation of the history ID 330 is performed by Mr. Suzuki to perform a print operation of a document considered to be a break in the operation. Cluster 121
Is divided into another cluster.

The cluster retrieving section 9 performs a process of retrieving and displaying clusters stored in the cluster storing section 8 in response to a user request. For example, document A. doc
When searching for a cluster that includes access to, the cluster IDs 121 and 122 are found. Since a plurality of operation histories including access to documents other than doc are compiled, the document A. doc with document C.doc. It can be seen that ppt may be used.

The cluster created by the cluster creating unit 7 is simply a collection of clusters that are close in time, such as a task performed over a long time or another task occurring between tasks. In such a case, the history may be interrupted (for example, in the case of the above-mentioned "Another email was received while the report was being created, and the document creation work was suspended to create a reply email." The history of the document creation work performed before the mail reception and the history of the document creation work resumed after the reply mail creation are one set for the purpose of the user.)

The inter-cluster similarity determination unit 10 determines the similarity between clusters in the operation history divided by the cluster creation processing as described above, and the case creation unit 11 has a high similarity (that is, the similarity is determined for the same purpose). Clusters), and create a user operation sequence (case) that is semantically connected. The similarity between clusters is determined based on an empirical rule that “the same document is accessed many times during the same business”, whether or not the cluster includes similar document operations. Empirically, since the same document is updated many times during the same business, clusters that include update operations of the same document have higher similarities, such as increasing the similarity. The similarity is weighted accordingly.

FIG. 4 shows an example of a case in which the similarity between the clusters shown in FIG. 3 is determined, and similar clusters are connected. In FIG. 4, case ID (c1)
Is a number assigned to manage a cluster group connected as a case. For example, case ID3
Reference numeral 2 denotes a cluster ID 121, 122, and 124 that are collectively sorted in the order of occurrence time. Cluster 12
1 and 122 have the same document A.1. It is considered to be similar because an operation that refers to doc is included.
In this example, the similarity is determined between clusters performed by different users so that work performed jointly by multiple users can be summarized as a case (in this case, Mr. Suzuki and Mr. Sato) Clusters are grouped together in the same case). Mr. Suzuki's cluster of the cluster 124 does not include a document common to the cluster 121, but the document C.C. Since access to the ppt is included, it is determined that they are similar, and they are put together in the same case. The created case is stored in the case storage unit 12.

The case search unit 13 performs a process of searching and displaying the cases stored in the case storage unit 12 in response to a user request, similarly to the cluster search process. For example, document A. Searching for cases that include access to doc, case 32 is found. From the case,
As with the cluster, the document A. doc with document C.doc. It can be seen that ppt is sometimes used, but since a plurality of operation histories are included as compared with the case of a cluster, it can be seen from the number of occurrences of document access that there is higher certainty in the relationship between documents. Further, the document C.1 is obtained from the operation occurrence pattern. The relationship that was not clear in the cluster, such as the relationship that Mr. Suzuki updated regarding ppt and the relationship referred to by Mr. Sato, can be understood.

FIG. 5 is a diagram showing the configuration of the information system according to this embodiment.

The main memory 101 stores a program being executed and control data necessary for executing the program. The hard disk 102 stores a control program, management data, and operation history. Things. The CPU 103 controls the operation of the entire apparatus. When the system is started, the program stored in the hard disk 102 is stored in the main memory 1.
01 and control is performed according to the contents.

For the interaction with the user, the keyboard 104 is used as an input device, and the result of the input operation is displayed on the display 105. As means for outputting a document, a printer 106 is connected. Further, this information system is connected to a network via a network interface 107, and performs processing for operations input from the network, outputs processing results to the network, or mails documents via the network. / Fax transmission is also possible.

These units are connected by a system bus, and are controlled by an operating system controlled by the CPU 103. The functions of the above-described embodiment are realized as application programs that operate on the operating system.

Next, the processing contents of this information system will be described in detail.

FIG. 6 is a main flow of the information system.

When this information system is activated, it initializes the entire system (step A1) and starts accepting user operations (step A2).

When the user makes an operation request for creating, referring to, updating or outputting a document to the system (YES in step A3), the information system stores the requested operation content in the operation history storage unit 6. (Step A4 to Step A)
5).

If the request from the user is to set a processing parameter (YES in step A6), the information system displays a dialog as shown in FIG. 7 and sets the processing parameter (step A7). .

In FIG. 7, the cluster time range (d1) sets the maximum value of the temporal size of one cluster, and the history data is set to a cluster that does not exceed the time range set here. Divided. Number of cluster entries (d
2) is to determine the maximum value of the number of operation histories included in one cluster, and to prevent the cluster from becoming too large even when document access occurs in a burst. .

The delimiter operation (d3) determines the type of operation to be detected as an operation delimiter. In this example, a document creation / update operation and a document delete operation are set as cluster delimiters. I have.

If the user's request is a cluster search process (step A8), this information system
If there is history data stored in the operation history storage unit 6, a history analysis process of creating a cluster from the history data is performed (step A9), and a cluster matching the user's request is selected from the clusters stored in the cluster storage unit 8 (step A9). Perform search and display processing (step A10 to step A1)
1). FIG. 8 shows a processing flow of processing for creating a cluster from history data.

When the cluster creation process is started, the history data stored in the operation history storage unit 6 in the order of occurrence time is divided for each user who has performed an operation (step B1). Next, if there is an operation set as a delimiter operation in the history sequence divided for each user, the history sequence is further divided immediately after the operation (step B2).

For each of the divided result histories, the registered “cluster time range” indicates the time from the occurrence of the first operation history to the occurrence of the last operation history.
The history group is divided such that the number of entries in the operation history included in the cluster is equal to or less than the number of registered cluster entries (step B3). Then, the cluster I created by the processing up to this point
A record is added to the cluster storage unit 8 with D attached, and all the history data stored in the operation history storage unit 6 is deleted (step B4).

In this embodiment, since the history data once processed for cluster creation is moved to the cluster storage unit 8, the target of the cluster creation processing at the time of retrieval has occurred since the previous cluster creation processing. It is only history data, and the load on analysis processing is distributed.

FIG. 9 is an example of a screen for setting search parameters when a user makes a cluster search request. In this example, the user who has performed the operation, each of the documents included in the operation content, the operation content of the document, and the time at which the operation occurred (e1 to e4) are selectively designated as necessary, and the cluster is designated. Perform a search. If there are a plurality of parameter values that can be specified, AND and OR specifications and wildcards are specified.

FIG. 10 is an example of a screen displaying a search result of a cluster satisfying a parameter designated by the user.
In this example, “Mr. Suzuki” includes the operation history of “update or print”, and displays a list of clusters generated between “12/23 to 12/31”.

When the request from the user is a system termination process (YES in step A12), the information system saves the management data and the history data on the hard disk 102 and executes the termination process ( Step A13).

As described above, according to the information system of this embodiment, the enormous amount of operation history stored in the operation history storage unit 6 can be converted into a history of processes performed by the user for the same purpose with a small amount of processing. The cluster can be converted into a grouped cluster, and the user can efficiently search and refer to the cluster.

Next, in addition to the above-described cluster creation,
Further, a case in which similar clusters are linked to create a case will be described. In this case, in addition to the above-described processing, processing for comparing clusters to determine the degree of similarity between clusters,
A process of linking similar clusters to create a case is added.

FIG. 11 shows a flow of processing for determining the similarity between clusters.

In this case, the information system first initializes all the similarities between clusters stored in the cluster storage unit 8 to 0 (step C1). Next, one cluster is selected from the cluster group (step C2), and the degree of similarity between the cluster and all clusters generated within the range set in the “maximum case time range” is calculated (step C2). C3).

The similarity is calculated mainly by comparing the documents included in the cluster and the types of operations on the documents. FIG. 12 shows an example of a screen for setting parameters used for calculating the similarity between clusters.

In FIG. 12, the case time range (f1) sets the generation time range of clusters that can be connected as a case, as described above. By setting this time, an unnecessarily long case is prevented from being created, and the load of the comparison processing can be suppressed by limiting the comparison to only the clusters within the time range.

The similarity determination parameter (f2) is used to calculate the similarity between clusters. For example,
Using the values set in FIG. 12, two clusters are compared, and if a certain document is referred to by both clusters, the similarity is increased by 5 points, and a certain document is updated by both clusters. If so, the similarity is increased by 100 points. In order to prevent a cluster that is not closely related from being judged to have a high similarity, when a document referred to by one cluster is not accessed by another cluster, the similarity is conversely increased by 2 points. Decrease.

A cluster including a large number of document access histories often includes an access history common to many clusters, and in order to avoid an unnecessary increase in similarity, the similarity of the common history included in the cluster is reduced. Finally, the similarity between clusters calculated as the accumulation of degrees is finally divided by the number of histories included in the cluster to obtain the similarity between clusters.

The remaining two parameters set in FIG. 12 are used in the case creation processing later. The cluster connection threshold value (f3) is used for comparison with the similarity between clusters, and is used for determining whether to connect two clusters. The setting (f4) of whether or not to connect clusters of a plurality of users is used to control whether or not to limit the connected clusters to operations performed by the same user.

Then, the information system executes the above-described similarity calculation for all clusters (step C4).

FIG. 13 shows a flow of processing for creating a case by connecting clusters based on similarity.

First, the information system selects a cluster (the cluster A having the earliest occurrence time of the operation history at the head of the cluster) that has occurred most recently from the clusters stored in the cluster storage unit 8, The contents of the cluster are registered in a newly created case (referred to as case A) (step D1).

Next, from the remaining clusters in the cluster storage unit 8, clusters that have occurred within the “case time range” are selected in the order of the generation time from the generation time of the operation history at the head of case A (cluster B). ), It is determined whether or not to combine with cluster A or case A (step D).
2). If "do not link clusters of multiple users" is specified as a parameter of the case creation process, the selection as cluster B is limited to the cluster performed by the same user who performed the operation of cluster A. .

As a determination method, for example, the cluster A
If the similarity between the cluster and the cluster B is larger than the “cluster connection threshold value”, it is determined that the clusters can be connected.
Alternatively, by averaging the similarities between cluster B and all the clusters already included in the case,
"Cluster connection threshold"
If it is larger than this, it is determined that the clusters can be connected. Further, when averaging the similarity between the cluster B and all the clusters already included in the case, the difference between the occurrence time of the cluster in the case and the occurrence time of the cluster B may be used as the weight of the similarity. Useful. For example, a value obtained by dividing the difference between the occurrence times of clusters 1 and 2 by the “case time range” and multiplying the similarity between clusters 1 and 2 is used as the weighted similarity of clusters 1 and 2. . Then, when the weighted similarity is greater than the “cluster connection threshold value”, it is determined that clusters can be connected.

If it is determined that cluster B can be connected to case A (YES in step D2), cluster B is connected to case A (step D3). Further, cluster B is newly set as cluster A, and the process returns to step D2 to continue selection of the next cluster and determination of the possibility of combination.

On the other hand, if no cluster connectable to case A is found (NO in step D2),
Are sorted in chronological order, registered with the case storage unit 12 with a case ID, and the registered cluster is stored in the cluster storage unit 8 so that the cluster registered in case A is not included in other cases. (Step D
4).

Steps D1 to D are executed until the case determination for all clusters is completed.
Step 4 is repeated.

The user's history search operation when the case creation process is performed is executed on the case storage unit 12. Except that the search target and display contents are not clusters but cases, the processing contents are performed in the same manner as the above-described cluster search / display processing.

In the above-described method, the case creation process is performed every time a user requests a case search. Therefore, if the case search process is performed frequently, the case creation process is performed on a very small number of history clusters. This will result in a short case that is not well connected. To avoid this, when registering a case, the difference between the time when the case was created and the time when the first operation history registered in the case occurred is the "case time range"
For smaller cases, a process of returning to the cluster without registering as a case and returning to the cluster storage unit 8 may be performed.

As described above, according to the information system of this embodiment, the clusters stored in the cluster
It can be converted into a case that is a larger aggregate in which similar clusters are connected to each other, and the user can efficiently search and refer to this case.

[0064]

As described above in detail, according to the present invention, the characteristics of the operation performed on the document file can be used with a small analysis load without putting a burden on the user regarding the accumulation and analysis of the history. It is possible to create a meaningful history set.

Further, by displaying the created history set to the user, it becomes possible to detect the relationship between the document, the user, and the operation content which cannot be obtained by the direct search of the history.

Further, by determining the similarity using the cluster, the similarity can be determined with a small load, and the operation histories can be grouped.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of an information system according to an embodiment of the present invention.

FIG. 2 is an exemplary view illustrating operation history data recorded in an operation history storage unit according to the embodiment;

FIG. 3 is an exemplary view illustrating a cluster created by a history cluster creating unit according to the embodiment;

FIG. 4 is an exemplary view exemplifying a case created by a case creating unit according to the embodiment;

FIG. 5 is an exemplary view showing the configuration of the information system of the embodiment.

FIG. 6 is a main flow of the information system of the embodiment.

FIG. 7 is an exemplary view showing an example of a processing parameter setting screen used for cluster creation according to the embodiment;

FIG. 8 is a processing flow of processing for creating a cluster from history data according to the embodiment;

FIG. 9 is an exemplary view showing a screen for setting a search parameter when making a cluster search request according to the embodiment;

FIG. 10 is an exemplary view showing an example of a screen displaying search results of clusters satisfying designated parameters according to the embodiment;

FIG. 11 is a processing flow when determining the similarity between clusters according to the embodiment;

FIG. 12 is an exemplary view showing an example of a screen for setting parameters used for calculating similarity between clusters according to the embodiment;

FIG. 13 is a processing flow of creating a case by connecting clusters based on the similarity according to the embodiment;

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 input device 2 display device 3 output device 4 document management unit 5 device monitor unit 6 operation history storage unit 7 history cluster creation unit 8 cluster storage unit 9 cluster search unit 10 similarity between clusters Judgment unit 11 Case creation unit 12 Case storage unit 13 Case search unit

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Tatsuya Haraguchi 70 Yanagimachi, Saiwai-ku, Kawasaki-shi, Kanagawa F-term in Toshiba Yanagimachi Plant (reference) 5B075 ND03 ND20 NK04 NR10 NR12 PP02 PP03 PR03 PR04 PR06 UU06 5B082 DD04 FA11

Claims (20)

[Claims] 1. A document management system having a function of accumulating an operation history for a document file, wherein the accumulated operation history is classified according to a predetermined rule, and a cluster is an aggregate having the operation history as an element. A document management system, comprising: a cluster creation unit that creates a document. 2. The cluster creating means creates a cluster by dividing the accumulated operation history for each operator, and then dividing the operation history divided for each operator for each predetermined time range. 2. The document management system according to claim 1, wherein: 3. The method according to claim 1, wherein the cluster creation unit divides the accumulated operation histories for each operator, and divides the operation histories for each operator into a predetermined number to create a cluster. The document management system according to claim 1, wherein 4. The method according to claim 1, wherein the cluster creation unit divides the accumulated operation history for each operator, and then creates a cluster by dividing the operation history divided for each operator by using a specific operation as a delimiter. The document management system according to claim 1, wherein: 5. The document management system according to claim 1, wherein said cluster creating means further comprises a cluster setting means for setting a rule for classifying operation histories. 6. A cluster search means for detecting a cluster satisfying a specified search condition from among a plurality of clusters created by the cluster creation means, and a cluster display means for displaying clusters detected by the cluster search means. The method according to claim 1, further comprising:
The document management system according to 2, 3, 4 or 5. 7. The document according to claim 1, further comprising: a similarity calculating unit that calculates a similarity between clusters created by the cluster creating unit. Management system. 8. The method according to claim 1, wherein the similarity calculating means calculates a similarity between the clusters from attribute information including an operation target, an operator, an operation date and time, and an operation content of an operation history constituting the cluster. Item 7. The document management system according to Item 7. 9. A method according to claim 1, wherein said similarity calculating means sets a value obtained by subtracting the number of inconsistencies of documents appearing only in one of the clusters from the number of matches of documents appearing in two clusters as the similarity between clusters. The document management system according to claim 7, wherein 10. The similarity calculating means performs a weighting value for each combination of operations performed on the same document from two clusters and a weighting value for any document from only one of the clusters. Of the documents appearing in only one of the clusters weighted by the weighting value based on the number of matches of documents appearing in the two clusters weighted by the weighting value. The document management system according to claim 7, wherein a value obtained by subtracting the number is used as a similarity between clusters. 11. The method according to claim 9, wherein the similarity calculating means averages the calculated values by the number of operation histories included in the clusters to obtain a similarity between the clusters.
Or the document management system according to 10. 12. The method according to claim 9, wherein the similarity calculating unit sets a value obtained by weighting the calculated value in accordance with a temporal distance of a cluster occurrence time as a similarity between clusters. 10. The document management system according to item 10. 13. The document management system according to claim 10, further comprising weighting value setting means for setting a weighting value used by said similarity calculating means to calculate the similarity. 14. A case creating means for creating a case that is a series of operation history groups by combining clusters created by the cluster creating means using the similarity between clusters calculated by the similarity calculating means. 14. The document management system according to claim 7, further comprising: 15. The document management system according to claim 14, wherein said case creating means includes means for joining closest clusters having similarities exceeding a predetermined threshold value. . 16. The case creation means, when an average of similarities between a single cluster and each of the clusters included in the already combined cluster group exceeds a predetermined threshold, 15. A means for coupling said cluster to said cluster group.
Document management system described. 17. When calculating the average of the similarity between a single cluster and each of the clusters included in the already-combined cluster group, the case creating means calculates the similarity between the clusters by the time of occurrence of the cluster. 17. The document management system according to claim 16, wherein weighting is performed according to the distance. 18. The document according to claim 16, wherein the case creating means determines only a cluster generated within a predetermined time range from a single cluster as a similarity determination target. Management system. 19. The apparatus according to claim 15, further comprising a case setting means for setting a threshold value, a weight value, or a time range used by said case creating means to combine clusters. , 17 or 18. 20. A case search means for detecting a case satisfying a specified search condition from a plurality of cases created by the case creation means, and a case display means for displaying the cases detected by the case search means. 14. The method according to claim 14, further comprising:
The document management system according to 6, 17, 18 or 19.