JP4163671B2 - Information retrieval system having database operation management support function and database operation management method - Google Patents

Description

  The present invention relates to an information search system capable of performing an information search for a plurality of databases in which information collected from different data sources is registered, and particularly has a database operation management support function suitable for operation management of the database. The present invention relates to an information retrieval system and a database operation management method.

  In recent years, knowledge sharing systems for supporting sharing of knowledge information have been developed. This knowledge sharing system is for accumulating and managing knowledge information such as personal know-how in a database (knowledge database). In addition, an information search system that realizes efficient use of knowledge information stored in a database by combining a knowledge sharing system and a search function such as natural language search has been developed.

  In an information retrieval system including a knowledge sharing system, how to efficiently collect knowledge information is an important point. Therefore, in recent information retrieval systems, it is common for the collection module to collect external data sources (such as information on the Internet or in-house intranets) and store the collected information in a database (for example, patents). Reference 1).

  In this type of information search system, when there are a wide variety of data sources, an operation of allocating one database for each data source is often applied. Such an operation is performed by an operation management client (operation management client terminal) based on an operation of an administrator (management user). The information search system has a search server so that information stored in the database can be used by the user. The search server searches the database specified by the search client (search client terminal) according to the user's operation according to the search condition, and presents the search result to the user via the search client. The user refers to the search result from the search server.

  By the way, the wide range of data sources means that various knowledge information can be collected. However, from the viewpoint of system operation management by the administrator via the operation management client, the number of databases to be managed is increased. It means inviting. In this case, it is difficult for an administrator to efficiently manage these databases.

  Further, the data automatically collected in a large amount by the collecting module includes a large amount of data unnecessary for the user. However, it is difficult for the administrator to determine what is unnecessary for the user.

On the other hand, when the information that is the search result (information selected from the search result list) is presented to the user via the search client, the user's evaluation for the information is obtained by inputting the user's evaluation for the information. Thus, it has been proposed to store user evaluation information corresponding to each piece of information in the database (see, for example, Patent Document 2). This Patent Document 2 describes that when a search result list or information selected from the search result list is presented to a user, user evaluation information corresponding to the information is also presented. Patent Document 2 also describes the use of user evaluation information corresponding to the ranking of searched information.
JP 2003-303197 A (paragraph 0008, paragraph 0010) JP 2002-41524 A (paragraphs 0009 to 0012, paragraph 0033, paragraph 0034, paragraph 0040, paragraph 0041, FIG. 8)

  In the information search system described above, information that must be managed by an administrator when knowledge data is frequently collected from a plurality of data sources to a corresponding database. The number of cases becomes enormous. In this case, it is difficult for the operation management client to efficiently manage these databases, and it is also difficult to determine a database unnecessary for the user.

  Therefore, it is conceivable to use user evaluation information for each registered information in each database described in Patent Document 2 for this determination. However, in order to use the user evaluation information for each information, the administrator has to check all the registration information in each database, and there is a problem that the cost of operation management increases. Moreover, as the registration information in the database increases, information that is less frequently referenced by the user also increases. Such information has a problem in terms of reliability of the user evaluation information itself because the number of evaluations by the user is reduced. Therefore, it is difficult for an administrator to determine whether a corresponding database is unnecessary for a user simply by using user evaluation information for each registered information in the database and evaluating the usefulness for each registered information. is there.

  The present invention has been made in consideration of the above circumstances, and its purpose is to easily determine the database that is effectively used without the administrator checking the usefulness of all information in each database individually. Another object of the present invention is to provide an information search system and database operation management method having a database operation management support function.

An information search system having a database operation management support function according to one aspect of the present invention includes a plurality of databases in which information collected from different data sources is registered, and a search condition specified in a search request from a user. Searching data that matches the search condition from at least one of a plurality of databases, and presenting the search results to the user, and using each of the plurality of databases, including a history of searches by the search server Log storage means for storing a history of status, and based on the usage status history of each of the plurality of databases stored in the log storage means, generating statistical information on the usage status of each of the plurality of databases, Statistics / evaluation means for evaluating the usage status of each of the plurality of databases based on statistical information; Comprising a database information presenting means for presenting the evaluation results of a plurality of databases each usage by meter and evaluation unit to the administrator, the statistical and evaluation unit, for each database, for each of a plurality of types of evaluation items Obtaining evaluation points based on the statistical information, and means for calculating an evaluation value for each database based on the obtained evaluation points, the evaluation value is an evaluation point of the evaluation item for each evaluation item The value obtained by multiplying the value obtained by multiplying the corresponding weighting factor and the value obtained by multiplying the number of documents in the database by the logarithm of the average number of documents in all databases for each evaluation item. Features.

  In such a configuration, the usage history of each of the plurality of databases, including the search history by the search server, is stored in the log storage means. The statistics / evaluation unit generates statistical information on the usage status of each of the plurality of databases based on the history of usage status of each of the plurality of databases stored in the log storage unit. The statistics / evaluation means evaluates the usage status of each of the plurality of databases based on the statistical information of the usage status of each of the plurality of databases. This evaluation result is presented to the administrator by the database information presenting means. As a result, the administrator can effectively use any one of the plurality of databases based on the evaluation result of the usage situation of each of the plurality of databases presented without checking all the registration information in each of the plurality of databases. , It is easy to determine which is not being used effectively. Therefore, the operation management cost can be reduced. Here, the evaluation result may be configured to be presented as an ordered list such as the order of databases that are used effectively or the order of databases that are not used effectively. At this time, an evaluation value may be given.

  Search history information indicating whether or not the usage history history of each of the plurality of databases stored in the log storage means is a history for each registered information in each of the plurality of databases and matches the search condition. Update presence / absence information indicating the presence / absence of update, a score indicating the degree of matching with the search condition, access presence / absence information indicating whether or not the searched information is selected and referred to by the user, and the information referred to by the user A configuration including a plurality of histories selected from the evaluation information indicating the user's evaluation may be applied. In this configuration, a plurality of information including a search frequency corresponding to the search hit presence information, an update frequency corresponding to the update presence information, the score, an access frequency corresponding to the access presence information, and a user evaluation corresponding to the evaluation information Among the evaluation items, at least one evaluation item selected and designated by the administrator from among the evaluation items corresponding to the plurality of selected histories is received by the statistics / evaluation unit and stored in the log storage unit Based on the information corresponding to the designated evaluation item in the usage history of each of the plurality of databases, statistical information on the usage situation of each of the plurality of databases related to the designated evaluation item is generated. If the usage status of each of the plurality of databases is evaluated based on the statistical information, a database with a high search frequency is used. Evaluation items intended by the administrator, such as databases with a lot of information that is frequently updated, databases with a lot of information that gives high scores in searches, databases with a lot of information with high evaluations, etc. It is possible to present the evaluation result of the database usage situation along the way.

  Further, if the usage status of each of the plurality of databases is evaluated so that the evaluation result of the usage status of each of the plurality of databases represents a tendency of change in the time axis direction, for example, the access frequency increases. A change in time series can be reflected in the evaluation result of the usage status of each database, such as the access frequency is low or the access frequency is low.

  According to the present invention, based on the usage history of each of the plurality of databases including the search history, the usage status statistical information of each of the plurality of databases is generated, and the usage status statistics of each of the plurality of databases are generated. Based on the information, the usage situation of each of the plurality of databases is evaluated, and the evaluation result is presented to the administrator so that the administrator does not individually check the usefulness of all the information in each database. However, it is possible to easily determine the database that is effectively used, and to reduce the cost of operation management.

Embodiments of the present invention will be described below with reference to the drawings.
[First Embodiment]
FIG. 1 is a block diagram showing a configuration of an information search system according to the first embodiment of the present invention. The information search system shown in FIG. 1 includes a plurality of databases, for example, two databases 11A (#A) and 11B (#B), collection modules 12A and 12B, a plurality of search clients (search client terminals) 13, and a search The server 14 includes a statistics / evaluation module 15, a statistics / evaluation log database 16, a database information presentation module 17, and an operation management client (operation management client terminal) 18. The collection modules 12A and 12B, the search server 14, the statistics / evaluation module 15, and the database information presentation module 17 are realized by a computer (server computer) reading and executing a dedicated program. This program can be stored in advance in a computer-readable storage medium and distributed. Further, this program may be downloaded (distributed) via a network.

  Hereinafter, the operation in the information search system of FIG. 1 will be described with reference to the sequence chart of FIG. The databases 11A and 11B are provided corresponding to different external data sources 21A and 21B (sources of information on the Internet or an intranet in the company), respectively, and accumulate information collected from the corresponding external data sources. Used. It is assumed that the external data source 21A exists on the Internet, for example, and the external data source 21B exists on the company intranet, for example. It is also possible to employ a configuration in which a database is prepared for each Web server using a plurality of predetermined Web servers on the Internet as external data sources. In order to simplify the description, it is assumed that information collected from the external data sources 21A and 21B is document information.

  The collection module 12i (i = A, B) collects document information from the external data source 21i, and registers the collected document information in the database 11i (step S1). When the collection module 12i collects document information corresponding to document information already registered in the database 11i, the newly collected document information is updated information of the registered document information (that is, an updated document). Determine whether. This determination is made by comparing the size or update date / time of the document information. When the newly collected document information is updated information, the collection module 12i updates the already registered document information with the newly collected document information. In addition, the collection module 12i notifies the search server 14 that the registered document information in the database 11i has been updated (step S2). If document information already registered in the database 11i does not exist in the external data source 21i, the collection module 12i deletes the registered document information from the database 11i.

  When the collection module 12i completes the collection of the document information in the database 11i, the search server 14 notifies the statistics / evaluation module 15 of information indicating whether or not each document in the database 11i is updated together with the database name of the database 11i (Step S14). S3). The statistics / evaluation module 15 uses the information notified from the search server 14 indicating the presence / absence of update as log information, attaches date / time information indicating the date / time at that time, and logs information described later in the statistics / evaluation log database 16 Save in the table 160 (step S4).

  The search client 13 requests the search server 14 to perform a search according to a user operation (step S5). In response to the search request from the search client 13, the search server 14 performs a search process according to the search condition specified in the search request for the database 11i (i is A or B) specified in the search request (step S6). In this search process, the search server 14 calculates a score representing the degree of matching with the search condition for each document registered in the database 11i. The search server 14 stores the calculated score for each document in a score field corresponding to the document in the search result table 140 described below.

  FIG. 3 shows an exemplary data structure of the search result table 140. The search result table 140 represents a search hit field that holds the presence or absence of a hit in the search for each of the databases 11A and 11B and for each document registered in the databases 11A and 11B, and the degree of matching with the search conditions. It has a score field that holds a score (the relevance ratio of a document searched for a search condition) and an access field that holds the presence or absence of access (reference). The search hit field is set to a state indicating that there is a search when the corresponding document is searched. The score field holds the score of the corresponding document calculated by the search process executed in response to the search request from the user. The access field is set to a state indicating that there is an access when the corresponding document is accessed (referenced).

  The search result table 140 also includes a “useful” evaluation field and a “useless” evaluation field for holding user evaluation results for each of the databases 11A and 11B and for each document registered in the databases 11A and 11B. Have The “useful” evaluation field is set to a state indicating that when the user evaluates that the corresponding document is useful. Similarly, the “useless” evaluation field is set to a state indicating that when the user evaluates that the corresponding document is useless. Note that the framework of the search result table 140 changes according to the increase / decrease in the database and the increase / decrease in the documents in the database.

  When the search server 14 calculates a score for each document registered in the database 11i, a document that matches (hits) the search condition, for example, a document with a score exceeding a predetermined threshold value, is stored in the search result table 140. The search hit field corresponding to the document is set to a state indicating that there is a search.

  The search server 14 transmits a search result including a list (search document list) of documents that match (hit) the search condition to the search client 13 (step S7). This search document list includes the document name (title name) of each document that meets the search condition and the document summary. The order of the documents (document names) in the search document list is the descending order of the corresponding scores.

  When a search result is transmitted from the search server 14 to the search client 13, a search result screen including the search document list is displayed on the display monitor of the search client 13. When the user performs an operation of selecting a document name of a desired document from the search document list, a document reference request (access request) is transmitted from the search client 13 to the search server 14 (step S8). In response to this reference request, the search server 14 retrieves the document information of the selected document name from the database 11i and transmits it to the search client 13 (step S9). Then, a document browsing screen including the document information selected by the user is displayed on the display monitor of the search client 13. As a result, the user can refer to the document information selected by the user on the document browsing screen.

  When the search server 14 transmits a document selected from the search result document list by the user to the search client 13, the search server 14 sets the access field corresponding to the document in the search result table 140 to a state indicating presence of access (reference). .

  On the document browsing screen displayed on the display monitor of the search client 13, as described in Patent Document 2, an evaluation input area for inputting a user's evaluation is secured. In this evaluation input area, radio buttons for inputting an evaluation type are arranged. Here, in order to simplify the explanation, the radio button for inputting the evaluation type is a radio used to input that the document information displayed on the document browsing screen (that is, the document information referred to by the user) is useful. Assume that there are two types of buttons: a button and a radio button used to input that the document information is useless. This example is equivalent to a configuration in which the user selects one of two evaluation values. In addition, you may be the structure which makes a user select one from the evaluation value of three steps or more.

  Now, when the document browsing screen is displayed, the user selects one of the two types of radio buttons arranged in the evaluation input area of the document browsing screen, thereby displaying the document information displayed on the document browsing screen. A specific evaluation can be given to the radio button with respect to (that is, document information referred to by the user). As a result of this evaluation, that is, whether the evaluation indicating “useful” or the evaluation indicating “useless” is input, the search client 13 notifies the search server 14 (step S10). . Note that by selecting neither of the two types of radio buttons, it is possible to give an evaluation that the document information is neither “useful” nor “useless”.

  If an evaluation indicating “useful” is input, the search server 14 sets the “useful” evaluation field corresponding to the document information evaluated as “useful” in the search result table 140 to “useful”. Is set to a state indicating that it has been evaluated. Similarly, if an evaluation indicating “useless” is input, the search server 14 evaluates the “useless” evaluation corresponding to the document information evaluated as “useless” in the search result table 140. The field is set to a state indicating that it has been evaluated as “useless”.

  For example, every time a series of search processes corresponding to one search condition is completed, the search server 14 transmits information of the search result table 140 at that time to the statistics / evaluation module 15 (step S11). When the statistics / evaluation module 15 normally receives the information of the search result table 140, the search server 14 initializes the state of each field of the search result table 140 in the search server 14.

  When receiving the information of the search result table 140 transmitted from the search server 14, the statistics / evaluation module 15 uses the information as log information, attaches date / time information indicating the date / time at that time, and adds the statistics / evaluation log database 16. In the log information table 160 described below (step S12).

  FIG. 4 shows an exemplary data structure of the log information table 160. The data structure of the log information table 160 is almost the same as that of the search result table 140 of FIG. Here, in order to simplify the description, every time information indicating whether or not each document in the database 11i is updated or information in the search result table 140 is transmitted from the search server 14 to the statistics / evaluation log database 16, It is assumed that the log information table 160 is generated in the statistics / evaluation log database 16. That is, in this embodiment, the columns of the log information table 160 are accumulated in the statistics / evaluation log database 16 in time series order. The log information table 160 is different from the search result table 140 in that it has an update field for holding the presence / absence of update.

  Now, when the presentation of database evaluation information is requested by the operation of the administrator, the operation management client 18 notifies the database information presentation module 17 of a request for presentation of database evaluation information (step S13). Here, the operation management client 18 displays a message prompting the administrator to specify an evaluation item and an evaluation period (an evaluation target period) when the administrator requests presentation of database evaluation information. . Thereby, the administrator can specify the evaluation item and the evaluation period.

  Upon receiving the presentation request from the operation management client 18, the database information presentation module 17 notifies the statistics / evaluation module 15 of the database evaluation request (step S14). This database evaluation request includes information on designated evaluation items and evaluation periods. When the statistics / evaluation module 15 receives the database evaluation request from the database information presentation module 17, the evaluation value Ri for each database 11i is used for the evaluation item specified in the evaluation request, using the evaluation point for each evaluation item. Is calculated (step S15). Evaluation items applied in the present embodiment are “useful”, “useless”, “high search frequency”, “low search frequency”, “high update frequency”, “low update frequency”, “score” 10 items of “High”, “Low score”, “High access frequency”, and “Low access frequency”, and one or a plurality of items can be selectively designated. Note that some of the above 10 items are not necessarily prepared. Conversely, the evaluation items “high registration frequency” and “low registration frequency” of information on the database 11i and “high deletion frequency” and “low deletion frequency” of information from the database 11i may be added. Is possible. For this purpose, at the time of collecting information by the collecting module 12i, in addition to the information indicating whether or not each document has been updated, information on a document newly registered in the database 11i or information on a document deleted from the database 11i. It is necessary to notify the statistics / evaluation module 15 of the information.

  Hereinafter, calculation of the evaluation value Ri for each database 11i will be described. First, the statistics / evaluation module 15 reads the columns of the log information table 160 that is log information in the designated evaluation period in order from the top. If the designated evaluation item is “useful”, the statistic / evaluation module 15 refers to the “useful” evaluation field of the log information table 160, and for each document in each database 11 i, in the designated evaluation period. Count the number of times a document has been evaluated as “useful”. When this count value is accumulated for all documents in the database 11i, the evaluation points (statistical information) of the database 11i regarding the evaluation item “useful” can be acquired.

  Similarly, if the designated evaluation item is “useless”, the statistic / evaluation module 15 refers to the “useless” evaluation field of the log information table 160 and designates the designated evaluation for each document in each database 11i. Count the number of times the document has been evaluated as “useless” during the period. By accumulating this count value for all the documents in the database 11i, the evaluation point of the database 11i regarding the evaluation item “useless” can be acquired.

  Next, if the designated evaluation item is “high search frequency” or “low search frequency”, the statistics / evaluation module 15 refers to the search hit field of the log information table 160 for each search. In addition, for each database 11i, it is determined whether the search frequency is high or the search frequency is low. Here, if the ratio of hit documents in one search with respect to the total number of documents in the database 11i is equal to or greater than the first threshold, it is determined that the search frequency of the database 11i is high. If the ratio of hit documents is less than the first threshold, it is determined that the search frequency of the database 11i is low. Then, every time it is determined that the search frequency of the database 11i is high, the operation of incrementing the evaluation point “high search frequency” by 1 is executed for the column of the log information table 160 in the designated evaluation period. Thereby, the evaluation point of the said database 11i regarding evaluation item "search frequency is high" can be acquired. Similarly, every time it is determined that the search frequency of the database 11i is low, the operation of incrementing the evaluation point “low search frequency” by 1 is executed for the column of the log information table 160 in the designated evaluation period. Thereby, the evaluation point of the said database 11i regarding evaluation item "search frequency is low" can be acquired.

  Next, if the designated evaluation item is “update frequency is high” or “update frequency is low”, the statistics / evaluation module 15 refers to the update field of the log information table 160 when information is collected. Thus, for each document in each database 11i, the number of times the document has been updated in the designated evaluation period (number of updates) is counted. Each time the statistical / evaluation module 15 detects a document whose number of updates in the specified evaluation period is equal to or greater than the average number of updates, the statistics / evaluation module 15 increments the evaluation point indicating that the update frequency is high in the corresponding database 11i by one. Thereby, the evaluation point of the said database 11i regarding evaluation item "update frequency is high" can be acquired. As the average number of updates, a value obtained by dividing the total number of updates for each document in the database 11i by the total number of documents updated even once in the database 11i is used. Further, each time the statistical / evaluation module 15 detects a document whose number of updates in the specified evaluation period is less than the average number of updates, the statistics / evaluation module 15 increments the evaluation point indicating that the update frequency is low in the corresponding database 11i by one. Thereby, the evaluation point of the said database 11i regarding evaluation item "update frequency is low" can be acquired.

  Next, if the designated evaluation item is “score is high” or “score is low”, the statistic / evaluation module 15 refers to the score field of the log information table 160 for each document in each database 11i. The number of times that the document has a score equal to or higher than the second threshold in the designated evaluation period is counted. By accumulating the count values for all the documents in the database 11i, the evaluation points of the database 11i regarding the evaluation item “score is high” or “score is low” can be acquired.

  Next, when the designated evaluation item is “access frequency is high” or “access frequency is low”, the statistics / evaluation module 15 refers to the access field of the log information table 160 when information is collected. Thus, for each document in each database 11i, the number of times the document is accessed (referenced) in the designated evaluation period (number of accesses) is counted. Each time the statistics / evaluation module 15 detects a document whose number of accesses in the specified evaluation period is equal to or greater than the average number of accesses, the statistics / evaluation module 15 increments the evaluation point indicating that the access frequency is high in the corresponding database 11i by one. Thereby, the evaluation point of the said database 11i regarding evaluation item "access frequency is high" can be acquired. For the average access count, a value obtained by dividing the total access count for each document in the database 11i by the total number of documents accessed even once in the database 11i is used. Further, each time the statistical / evaluation module 15 detects a document whose access count in the specified evaluation period is less than the average access count, the statistics / evaluation module 15 increments an evaluation point indicating “access frequency is low” in the corresponding database 11 i by one. Thereby, the evaluation point of the said database 11i regarding evaluation item "access frequency is low" can be acquired.

When the statistics / evaluation module 15 obtains evaluation points for the designated evaluation item for each database 11i (i = A, B), the evaluation value (statistical evaluation value) Ri for each database 11i is obtained based on the evaluation points. calculate. This evaluation value Ri is given by the following formula (1)
Ri = W1f1 + W2f2 + W3f3... + Wnfn (1)
It is represented by Here, n is the number of evaluation items to be applied. In this embodiment, n = 4, and W1f1, W2f2, W3f3 and Wnfn (= W4f4) are “high access frequency”, “low access frequency”, “useful”, and “useless”, respectively. The evaluation value for each evaluation item is represented, and W1, W2, W3, and W4 represent the weights of the corresponding evaluation items.

In the formula (1), fn is
fn = number of documents in database 11i × log (average number of documents in all databases) (2)
It is represented by If the weights W1, W2, W3, and W4 are represented by W, the weight W is W = (evaluation point × evaluation coefficient) / f (x) (3)
It is represented by Here, f (x) represents the number of documents normalized with respect to all databases, for example, the average number of documents in one database. The evaluation coefficient is a value specific to the evaluation item.

  FIG. 5 shows an example of 10 types of evaluation items applied in the present embodiment and evaluation coefficients predetermined for each evaluation item.

  The statistics / evaluation module 15 transmits the evaluation value Ri for each database 11i calculated as described above to the database information presentation module 17 (step S16). Upon receiving the evaluation value Ri for each database 11i transmitted by the statistics / evaluation module 15, the database information presentation module 17 presents the evaluation value Ri for each database 11i to the administrator via the operation management client 18 (step S17). At this time, if the evaluation values Ri for each database 11i are sorted and presented in the order of high or low evaluation value, the administrator ranks the databases 11i that are used effectively or ranks that are not used effectively. Is easily visible. Also, it is preferable that the administrator can specify whether the evaluation values are presented in order of high or low. From the evaluation value Ri for each database 11i, that is, the statistical evaluation result of the usage situation for each database 11i, the administrator who is the user of the operation management client 18 is the database that is used effectively or is used effectively. Can determine which database is not. The evaluation value Ri for each database 11i is a measure for deleting a database that is not being used effectively. Thereby, the cost of operation management can be reduced.

  Of course, the database information presentation module 17 or the operation management client 18 automatically determines the database that is used effectively or the database that is not used effectively by comparing the evaluation value Ri for each database 11i with a threshold value. Is also possible.

  As is apparent, in the above-described example, a database having a higher “access frequency” evaluation value and a “useful” evaluation value, and a lower absolute value of a “less access” evaluation value and a “useless” evaluation value. The evaluation value Ri increases as 11i increases.

  In the first embodiment, four types of evaluation items “high access frequency”, “low access frequency”, “useful”, and “useless” are applied to the calculation of the evaluation value Ri of the database 11i. ing. However, the arbitrary evaluation items shown in FIG. 4 can be used alone or in combination. For example, when “high score” is designated as the evaluation item from the operation management client 18, that is, when presentation of evaluation order for “database with many data having high scores” is designated, the statistics / evaluation module 15 Focusing on only “high score”, the evaluation value Ri for each database 11i is calculated. The evaluation value Ri for each database 11 i is presented to the operation management client 18 by the statistics / evaluation log database 16. Thereby, the administrator can discriminate the database that is used effectively from the evaluation result (evaluation value Ri) for each database 11i evaluated from the viewpoint of "a database with many data with high scores". Further, when the evaluation value Ri for each database 11i is calculated by applying, for example, “high access frequency”, “low access frequency”, “high update frequency”, and “low update frequency” as evaluation items. For example, it can be used as a guideline for deleting a database with low access frequency and low update frequency (that is, many old documents).

  In the first embodiment, a configuration in which the evaluation value Ri for each database 11 i is transmitted from the database information presentation module 17 to the operation management client 18 in response to a request from the operation management client 18 is applied. However, the statistical / evaluation module 15 calculates the evaluation value Ri of each database 11i by applying the evaluation items designated in advance by the operation management client 18 at predetermined timings, for example, at regular intervals. The evaluation value Ri for each database 11i can be periodically transmitted from the database information presentation module 17 to the operation management client 18.

  Further, in the first embodiment, the period (evaluation period) of log information used for calculating the evaluation value Ri can be designated from the operation management client 18. The significance of designation of this evaluation period will be described below. For example, the evaluation value Ri based on the log information in “all periods” from when the database 11i is constructed to the present is different from the evaluation value Ri based on the log information in “a specific period” within the entire period.

  Therefore, in the first embodiment, the evaluation value Ri is calculated in consideration of the tendency of the log information to change in the time axis direction. Here, it is assumed that an evaluation value Ri for “useful” is calculated for a certain database 11i.

  First, the number of documents evaluated as “useful” in the X years (X is an integer of 2 or more, for example) from the creation of the database 11i to the present is 3,000. Assume that the number of documents evaluated as “useful” is 2000, and the number of documents normalized to all databases (for example, the average number of documents in one database) is 2000.

In this case, if the change tendency in the time axis direction is not taken into consideration, the evaluation value Ri is Ri = (the number of documents evaluated as “useful” 3000 × evaluation coefficient 3 of “useful”)
/ 2000 documents normalized to all databases
= 4.5
It becomes.

On the other hand, considering the trend of change in the time axis direction, focusing on the most recent year of X years,
Ri = (the number of documents evaluated as “useful” 2000 × evaluation coefficient 3 of “useful”)
/ 2000 documents normalized to all databases
= 3
It becomes.

  This difference indicates that the degree that documents registered in the database 11i are evaluated as “useful” tends to decrease compared to before that in the last one year.

Here, the advantages of limiting the evaluation period to “a specific period” in consideration of the tendency of change in the time axis direction will be described.
First, the number of documents registered in the databases 11A and 11B in FIG. 1 is the same, and the access frequency is high. Database 11A: Access frequency is high until one year ago Database 11B: Until the past one year It is assumed that the access frequency is low, but there is a tendency that the access frequency is high in the recent one year.

Here, as a result of calculating the evaluation values RA and RB of the databases 11A and 11B in the period from the creation of the databases 11A and 11B to the present without considering the change tendency in the time axis direction,
Evaluation value RA of database 11A = 5
Evaluation value RB = 4 of database 11B
Is acquired. In this case, since the evaluation value RA> the evaluation value RB, the database 11A is determined as a more effective database.

However, depending on the administrator, it may be more effective to discriminate a database having a high access frequency from the evaluation values of the last one year than the evaluation values in the period from the creation of the databases 11A and 11B to the present. Therefore, as a result of calculating the evaluation values RA and RB with the latest one year as the evaluation period, the evaluation value RA of the database 11A, for example, is 4.2 from the above tendency.
Evaluation value RB of database 11B = 4.8
Suppose that In this case, when the evaluation is focused on the latest one year, a result that the database 11B is more effective is obtained. Thus, by making it possible to designate the evaluation period from the operation management client 18 in accordance with the operation of the administrator, it becomes possible to use a more effective database for each administrator. It is also possible to calculate an evaluation value for each fixed period, such as a year unit or a half year unit, and present the change tendency of the evaluation value to the administrator.

[Second Embodiment]
In the first embodiment, the evaluation value Ri for each database 11 i is presented to the administrator by the database information presentation module 17 via the operation management client 18. The administrator can use the presented evaluation value Ri for each database 11i as a guideline when deleting a database that is not being used effectively. However, instead of presenting the evaluation value Ri for each database 11i to the administrator, it is possible to change the automatic update frequency of the database from the evaluation value Ri for each database 11i.

  Therefore, a second embodiment of the present invention that realizes automatic database update will be described with reference to the drawings. FIG. 6 is a block diagram showing a configuration of an information search system according to the second embodiment of the present invention. In FIG. 6, components equivalent to those in FIG.

  6 is different from the system of FIG. 1 in that an automatic update module 19 is used in place of the database information presentation module 17 and the operation management client 18. The automatic update module 19 calculates an automatic collection condition of information by the collection module 12i for each database 11i, for example, an automatic update frequency (automatic collection frequency) from the evaluation value Ri for each database 11i calculated by the statistics / evaluation module 15. Decide (change). For this purpose, the automatic update module 19 receives an evaluation value Ri for each database 11 i from the statistics / evaluation module 15. Then, the automatic update module 19 collects from the evaluation value Ri for each database 11i so as to reduce the automatic update frequency for a database with low access frequency, for example, and increase the automatic update frequency for a database with high access frequency. Controls modules 12A and 12B. Thus, the user of the search client 13 can always search for the latest valid data, and the operation management cost can be further reduced. In addition, it is also possible to control the collection modules 12A and 12B so that the collection target (collection range) is expanded or the collection target is narrowed according to the access frequency.

  Further, the database information presentation module 17 and the operation management client 18 shown in FIG. 1 are added to the system of FIG. 6, and the evaluation value Ri for each database 11 i is managed by the database information presentation module 17 via the operation management client 18. You may make it show to a person.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, you may combine suitably the component covering different embodiment.

1 is a block diagram showing a configuration of an information search system according to a first embodiment of the present invention. The sequence chart which shows the procedure of the operation | movement in the embodiment. The figure which shows the data structure example of the search result table 140 in FIG. The figure which shows the example of a data structure of the log information table 160 in FIG. The figure which shows an example of 10 types of evaluation items applied in the embodiment, and an evaluation coefficient predetermined for each evaluation item. The block diagram which shows the structure of the information search system which concerns on the 2nd Embodiment of this invention.

Explanation of symbols

  11A, 11B ... database, 12A, 12B ... collection module, 13 ... search client, 14 ... search server, 15 ... statistics / evaluation module, 16 ... statistics / evaluation log database (log storage means), 17 ... database information presentation module, 18 ... Operation management client, 19 ... Automatic update module, 21A, 21B ... External data source, 140 ... Search result table, 160 ... Log information table.

Claims (10)

Multiple databases with information collected from different data sources,
A search server that searches for data matching the search condition from at least one of the plurality of databases according to the search condition specified in the search request from the user, and presents the search result to the user;
Log storage means for storing a history of usage of each of the plurality of databases, including a history of searches by the search server;
Based on the history of the usage status of each of the plurality of databases stored in the log storage means, generates statistical information on the usage status of each of the plurality of databases, and based on the statistical information, Statistics / evaluation means to evaluate usage,
Database information presenting means for presenting an evaluation result of the usage status of each of the plurality of databases by the statistics / evaluation means to an administrator ;
The statistics / evaluation means includes means for obtaining an evaluation point based on the statistical information for each of a plurality of types of evaluation items for each database and calculating an evaluation value for each database based on the obtained evaluation points. And
The evaluation value is obtained by multiplying, for each evaluation item, a weight coefficient corresponding to the evaluation point of the evaluation item and a value obtained by multiplying the number of documents in the database by the logarithm of the average number of documents in all databases. An information search system having a database operation management support function, characterized in that a value is obtained by adding a value for each evaluation item . The usage history history of each of the plurality of databases stored in the log storage means is a history for each registered information in each of the plurality of databases, and search hit presence / absence information indicating whether or not a search condition is satisfied, Update presence / absence information indicating the presence / absence of update, a score indicating the degree of matching with the search condition, access presence / absence information indicating whether or not the searched information is selected and referred to by the user, and the user referred to by the user Including multiple histories selected from the evaluation information indicating the evaluation of
The statistics / evaluation means includes a search frequency corresponding to the search hit presence information, an update frequency corresponding to the update presence information, the score, an access frequency corresponding to the access presence information, and a user evaluation corresponding to the evaluation information And receiving at least one evaluation item selected and designated by an administrator from among the evaluation items corresponding to the selected plurality of histories. Based on the information corresponding to the designated evaluation item in the history, the usage information of each of the plurality of databases related to the designated evaluation item is generated, and the plurality of databases is generated based on the statistical information. The information retrieval system having a database operation management support function according to claim 1, wherein each usage situation is evaluated. Stem.   2. The statistics / evaluation unit evaluates the usage status of each of the plurality of databases so that the evaluation result of the usage status of each of the plurality of databases represents a tendency of change in a time axis direction. An information search system having the described database operation management support function.   The system further comprises an automatic deletion unit that automatically deletes a database having a low evaluation among the plurality of databases based on an evaluation result of a usage situation of each of the plurality of databases by the statistics / evaluation unit. An information retrieval system having the database operation management support function according to Item 1. The system further comprises means for urging an administrator to delete a database having a low evaluation from the plurality of databases based on the evaluation result of the usage status of each of the plurality of databases by the statistics / evaluation means. An information retrieval system having the database operation management support function according to claim 1 . Information collecting means for collecting information from the different data sources according to set information collecting conditions and registering the collected information in the database corresponding to the data sources;
2. The database operation management support according to claim 1 , further comprising: an automatic update unit that changes the information collection condition based on an evaluation result of a usage situation of each of the plurality of databases by the statistics / evaluation unit. Information retrieval system with functions. Information collecting means for collecting information from the different data sources according to set information collecting conditions and registering the collected information in the database corresponding to the data sources;
The database operation management according to claim 1 , further comprising means for prompting an administrator to change the information collection condition based on an evaluation result of the usage status of each of the plurality of databases by the statistics / evaluation means. An information retrieval system having a support function. Multiple databases with information collected from different data sources,
A search server that searches for data matching the search condition from at least one of the plurality of databases according to the search condition specified in the search request from the user, and presents the search result to the user;
Log storage means for storing a history of usage of each of the plurality of databases, including a history of searches by the search server;
Based on the history of the usage status of each of the plurality of databases stored in the log storage means, generates statistical information on the usage status of each of the plurality of databases, and based on the statistical information, Statistics / evaluation means to evaluate usage,
Information collecting means for collecting information from the different data sources according to set information collecting conditions and registering the collected information in the database corresponding to the data sources;
Automatic update means for changing the information collection condition based on the evaluation results of the usage status of each of the plurality of databases by the statistics / evaluation means ,
The statistics / evaluation means includes means for obtaining an evaluation point based on the statistical information for each of a plurality of types of evaluation items for each database and calculating an evaluation value for each database based on the obtained evaluation points. And
The evaluation value is obtained by multiplying, for each evaluation item, a weight coefficient corresponding to the evaluation point of the evaluation item and a value obtained by multiplying the number of documents in the database by the logarithm of the average number of documents in all databases. An information search system having a database operation management support function, characterized in that a value is obtained by adding a value for each evaluation item . In accordance with a search condition specified by a search request from a user, data that matches the search condition is searched from at least one of a plurality of databases in which information collected from different data sources is registered, and the search result is In an information search system presented to a user, a database operation management method for managing operations of the plurality of databases,
Each time the search according to the search condition specified by the search request from the user is performed by the statistics / evaluation means, the information acquired by the search is logged as a part of the history of the usage status of the database subjected to the search. Storing in the storage means;
The statistical / evaluation means , based on a history of usage status of each of the plurality of databases stored in the log storage means, generating statistical information of usage status of each of the plurality of databases;
The statistic / evaluation means evaluating the usage status of each of the plurality of databases based on the generated statistical information of the usage status of each of the plurality of databases;
Database information presenting means comprising the step of presenting to the administrator the evaluation results of the usage status of each of the plurality of databases ,
The step of evaluating by the statistical / evaluation means obtains an evaluation point based on the statistical information for each of a plurality of types of evaluation items for each database, and calculates an evaluation value for each database based on the obtained evaluation points. Comprising the steps of:
The evaluation value is obtained by multiplying, for each evaluation item, a weight coefficient corresponding to the evaluation point of the evaluation item and a value obtained by multiplying the number of documents in the database by the logarithm of the average number of documents in all databases. A database operation management method characterized by being a value obtained by adding a value for each evaluation item . The usage history history of each of the plurality of databases is a history for each registered information in each of the plurality of databases, and represents search hit presence / absence information indicating whether or not the search condition is met, and the degree of matching to the search condition. Including a plurality of histories selected from the score, the access presence / absence information indicating whether or not the searched information is selected and referred to by the user, and the evaluation information indicating the user's evaluation with respect to the information referred to by the user ,
Among the plurality of evaluation items including the search frequency corresponding to the search hit presence / absence information, the score, the access frequency corresponding to the access presence / absence information, and the user evaluation corresponding to the evaluation information, When at least one evaluation item is selected and specified by the administrator from among the corresponding evaluation items, each of the plurality of databases related to the specified evaluation item based on information corresponding to the specified evaluation item The database operation management method according to claim 9, wherein statistical information on the usage status of the database is generated, and usage status of each of the plurality of databases is evaluated based on the statistical information.

Images