JPH10326281A - Method and device for information retrieval and management - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve retrieval efficiency and to enable high quality retrieval by retrieving a similar category from categories that a directory service provides when an example is given. SOLUTION: After collecting information, an information storage part 13 stores it in an index information data base 13a. An information retrieval part 15, once given an index example, performs a retrieving process and presents a provider 17 for similar directory service and a category. Consequently, when an index example is obtained actually from a user, the information retrieval part 15 calculates similarity between the index example as an example and category information stored in the index information data base 13a and displays the provider 17 for similar directory service and the category. Therefore, new information can be found according to the category to improve the efficiency, and collected information is already categorized, so the information can easily be rearranged.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an information search / management method and an information search / management apparatus suitable for a directory service in the Internet or a database system.

[0002]

2. Description of the Related Art As shown in FIG. 13, an information retrieval apparatus used in a conventional retrieval service (retrieval service) stores target information collected by an information collection unit 111 of an information retrieval apparatus 101 in an information storage unit. In addition to recording and accumulating the information in the content information database 113a via the index 113, the index assigning unit 117 assigns an index using a natural language analysis technique, and records and accumulates the index in the index information database 117a.

At the time of information retrieval, an information retrieval unit 115
In the content information database 113, a search is performed from the index information recorded and stored in the index information database 117a using a keyword or a combination of a plurality of keywords.
The corresponding information is found from the information recorded and stored in a. In such a search service, since pattern matching is performed on the character surface, there are the following problems. (1) Even if the meaning is the same, if different expressions are made, they are not searched. (2) Even if the information is completely unrelated, if the character face is the same, it becomes a search target.

[0004] On the other hand, volunteers and specific companies,
Information in a specific field is organized and classified by an organization or the like, and is provided as an index service (directory service). In these services, since classification and sorting work is performed manually, the service contents are often very detailed and high quality.

However, this has the following problems. (1) Since it is a manual operation, the target field is limited, and it is difficult to find an organization providing a directory service useful for the user.

[0006] (2) Since the point of view of organization differs for each directory service, it is difficult to understand the difference in the category of the directory service, and it is necessary to master the use of the different directory services sufficiently.

[0007]

As described above, as described above,
In the past, trial and error was required to find a directory service of interest. In addition, although it is possible to find a related directory service by specifying a related keyword in a conventional search service, the quality of information output due to the drawbacks of the search service as described above. Is low, and it often takes time and effort to find desired information.

The present invention has been made in view of the above problems, and has as its object to efficiently search for information on the Internet or an intranet, which is an in-house network, which has exploded in recent years, and to manage the searched information. And In particular, the present invention increases the search efficiency by providing a method of searching for an index service in addition to the conventional search service.
It is an object of the present invention to provide an information search / management method and an information search / management device that enable high-quality search.

[0009]

In order to achieve the above-mentioned object, an invention according to claim 1 of the present invention is an information search / management method for presenting a predetermined directory service provider and a category. The gist of the present invention is to search for similar categories from the categories provided by the directory service by giving examples.

Thus, according to the present invention, by performing index service search in addition to the conventional search service, more efficient information search can be performed.

The invention according to claim 2 is an information search / management method for presenting an optimal directory service provider and category when a plurality of directory services are provided, respectively, Collecting, recording and storing category information from a plurality of providers, and, when an example is given, calculating the similarity from the example and the stored category information, and selecting an optimal directory service from the similarity. And a step of presenting the category and the category.

The invention according to claim 3 is an information search / management apparatus for presenting a predetermined directory service provider and a category when a directory service is provided from a plurality of providers, respectively. Information collecting means for collecting category information from the provider, information storing means for recording and storing category information collected by the information collecting means, and when an example is given, the example is stored in the information storing means. The gist is to calculate the similarity from the category information, and to have a predetermined directory service provider and an information search means for presenting the category from the similarity.

As a result, it is possible to search for a directory service that is close to the information collected and arranged by the user, and to search for similar information with high quality.

[0014] Preferably, the information storage means according to claim 3 is provided with a category record table and an information entry record table.

Preferably, the degree of similarity according to the third aspect is an index represented by the inclusion rate and the content rate. That is,
By showing the similarity (inclusion rate, content rate) between a given index example and a category, the difference between categories can be clarified.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an embodiment of an information search device 1 to which an information search / management method according to the present invention is applied. As shown in FIG. 1, the information search device 1 includes an information collection unit 11, an information storage unit 13, an index information database 13a, and an information search unit 15.

The directory service provider 17 provides the information retrieval device 1 with a category representing a concept and an information entry representing the location of actual information. Certain categories may include categories and information entries that include subordinate concepts. Therefore,
The directory service as a whole has a tree structure indicating a hierarchy of concepts. A user using the directory service can find the location of the desired information by tracing the hierarchy of the desired category.

The information collecting unit 11 collects category information from a plurality of providers that provide directory services A, B,. Upon collection, the information on the categories and information entries provided by the directory service and their hierarchical relations are efficiently obtained by sequentially tracing the lower concepts from the highest concept (root category) of each directory service. can do.

The information storage unit 13 stores the collected information in a category record table (see FIG. 5) and an information entry record table (see FIG. 6) of the index information database 13a after performing a collecting process described later. . Note that the above information collection and accumulation operation is executed in advance before an information search request is issued.

When an index example is given, the information search unit 15 performs a search process described later and presents similar directory service providers and categories. This allows
When an index example is actually shown by the user, the information search unit 15 calculates the similarity (inclusion rate, content rate) from the index example as an example and the category information stored in the index information database 13a of the information storage unit 13. , Offer similar directory service providers and categories. Based on this information, the user can actually check the directory service.

Next, calculation of the similarity between an example and a category in the information search / management method of the present embodiment will be described. First, the examples and directories presented by the user
An index represented by the inclusion rate and the content rate is used for calculating the similarity with the category information collected from the service provider. Assuming that a set of information entries included in a category shown as an example is K and a set of information entries included in a certain collected category is N, the inclusion rate and the content rate are defined as follows. Inclusion: What percentage of K is included in N? Content: What percentage of N is common with K?

FIG. 2 shows the inclusion rate and content rate between categories,
That is, it is a diagram showing the relationship between the example set (K) and the target set (N). As is clear from FIG. 2, it is most ideal that the same category as the example is found (case C2). In the case C2, both the inclusion rate and the content rate are naturally 100%. However, since the purpose is to actually know the information entry which is not in the example, the content rate is 100% and the content rate is smaller than 100% (case C1). May be smaller than 100% (case C4) and is also desirable.

When the content is too small, the category is often very large. Therefore, if the content is the same, the content is preferably as large as possible.

Further, in the present definition, even when a category has subcategories and is nested, the calculation is performed only by the number of information entries, ignoring the subcategories.
In addition to the above calculation of the index, a method of including the number of subcategories included in the category in the calculation is also conceivable. In this case, for the categories having the same information entry, those having a subcategory and those having no subcategory have different numerical values for the index.

In addition to the above-mentioned indices, it is possible to define and use indices. For example, it is possible to create an index based on information such as the percentage of the information provided by the provider and the depth of the target category, and use it for a search.

Next, an information collection algorithm will be described with reference to FIGS. FIG. 3 shows an information collection algorithm (collection processing), and FIG. 4 shows an information collection algorithm (category collection processing).

Referring to FIG. 3, when collecting information,
First, in step S11, the category record table (see FIG. 5) and the information entry record table (see FIG. 6) are initialized. Next, the creation of a provider list for the target provider is performed for all providers together with the collection of categories for each category of each provider (step S17) (steps S13, S15, and S).
17).

Next, a category collection process will be described with reference to FIG. First, in step S21, the given category is recorded in the category record table, and it is further confirmed whether or not the given category is an information entry. Here, when it is confirmed that the category is not an information entry, the process proceeds to step S25, and a category collection process is performed on all subcategories of the given category. When it is confirmed that the category is an information entry, the process proceeds to step S27, and the given information is recorded in the information entry record table. Thereby, based on the provider list given in advance, the category can be recursively traced up to the part where the information entry is stored, and can be recorded together with the provider name and the relation with the upper and lower ranks.

Next, the configuration of an information storage table for storing collected information will be described with reference to FIGS. FIG. 5 shows an example of the category record table, and FIG. 6 shows an example of the information entry record table. The category record table is prepared for storing category information, and the information entry record table is prepared for quickly searching for the same entry as the input information entry. At this time, the information entry is hashed. Stored.

Each entry of the category record table of FIG. 5 has a “number field” indicating the serial number of the entry,
"Type field" indicating the category type, "category / provider / information entry name field" indicating the name of the category or provider or information entry, "upper category field" indicating the upper category of the category, "lower level" indicating the lower category Category Field "
Consists of The first entry in FIG. 5 is the provider (DS1)
Shows the information. Type is provider, name DS1,
Since this is the highest category, there is no higher category, indicating that the category has second and fourth entries as lower categories. The second entry is a sub-category (DS1-C1) of the provider shown in the first entry, the upper category is the first entry, and indicates that the lower category has information of the third, fourth, and fifth entries. ing. Hereinafter, a state in which similar information is stored is shown. FIG. 6 shows an example of the information entry record table. The table implements an information name and a category record table entry number as a hash table. Since the hash table uses the information name as a key, the entry number of the category record table can be searched from the information name by using this table. In the search of the category record table from the actual information name, the entry number of the information entry record table is calculated using the hash function from the information name, and the storage location of the category record table can be further known from the entry.

Next, an information retrieval algorithm will be described with reference to FIGS. FIG. 7 shows a similar category search algorithm, and FIG. 8 shows an optimal similar category search algorithm.

Referring to FIG. 7, in the similar category search process, the information example given in step S31 is checked in the information entry record table. Next, in step S33, the already-checked information entry next to the current information entry is taken out. When the information entry is completed in step S35, the search processing is terminated. Otherwise, the process proceeds to step S37, and the next higher order is performed. Retrieve a category. When the next upper category is extracted in step S37, if there is no corresponding upper category, or if the calculation has already been performed, the process returns to step S33; otherwise, the process proceeds to step S41, and the optimum After performing similar category search processing, step S33
Return to

Referring to FIG. 8, in the optimal similar category search process, first, in step S51, the inclusion rate and content rate of a given category are calculated and recorded in a table. In step S53, if the category is equal to the root category, the search process is terminated. Otherwise, the process proceeds to step S55 to check whether the inclusion rate is 100%. If it is 100%, the search process is terminated. Otherwise, the process proceeds to step S57, where the magnitude relation between the coverage of the upper category and the coverage of the given category is checked. If the inclusion rate of the upper category is larger than the inclusion rate of the given category, the search processing is terminated. Otherwise, the process proceeds to step S59, where an optimal similar category search of the upper category is performed.

As described above, in the present embodiment, the category including the corresponding information entry is found from the category record table based on the example information entry input from the information entry table, and the category record table is traced upward. Meanwhile, the similarity index represented by the inclusion rate and the content rate is calculated.

Here, a case has been described in which a category is searched from a fictional directory service. However, on the Internet, the information entry is actually a URL (Universal Rseso).
urceLocator), and the directory service provider is a hot list published on a personal homepage or link information of hobby information.

FIG. 9 shows a service provider (DS1, DS
2, DS3) shows the result of storing the result of collecting category information. From among these, (C
(1, C2, C4).

First, an entry including an example is checked on the information entry storage table. In FIG. 10, the corresponding entry is shaded. Next, as shown in FIG. 11, attention is paid to the sub-subcategory including the checked entry, and the similarity index is calculated. Further, the similarity index calculation is repeated in the order of the sub-subcategory, the sub-category, and the category until the inclusion rate does not increase or reaches the root category.

After the above calculation is completed, a list of categories having the highest inclusion rate and high content rate is shown to the user together with their indexes (FIG. 12). It is also possible to show higher and lower concepts according to the user's request. Prioritize content or inclusion based on user preference,
It is also possible to prioritize and present them to the user by weighting them.

In the above embodiment, the case where the present invention is applied to a directory service in the Internet or a database system has been described as an example. However, the present invention is not limited to this. For example, another similar category is searched. Applications, specifically intranets, information management, indexing services, database management,
The present invention can be applied to appropriate fields such as information collection and search engines.

[0040]

As described above, according to the present invention, it is possible to efficiently search for similar categories by giving examples to information classified into categories. URL
Based on the information, it is possible to efficiently find directory services and their categories arranged in similar categories. As a result, new information can be found based on the category, and the efficiency is improved. Further, since the collected information is already categorized, the information can be easily organized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of an information search / management apparatus to which an information search / management method according to the present invention is applied.

FIG. 2 is a diagram showing a relationship between an inclusion rate and a content rate and categories.

FIG. 3 is a flowchart illustrating an information collection algorithm (collection processing).

FIG. 4 is a flowchart illustrating an information collection algorithm (category collection processing).

FIG. 5 is a diagram showing a configuration of a category recording table.

FIG. 6 is a diagram showing a configuration of an information entry record table.

FIG. 7 is a flowchart illustrating an information search algorithm (similar category search processing).

FIG. 8 is a flowchart illustrating an information search algorithm (optimal similar category search processing).

FIG. 9 is a diagram illustrating a storage state of collected category information.

FIG. 10 shows C1, C2, and C3 from stored category information.
It is a figure for explaining search of the category containing C4.

FIG. 11 is a diagram illustrating a process of calculating a similarity index.

FIG. 12 is a diagram showing search results of categories.

FIG. 13 is a block diagram showing a configuration of a conventional information search device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Information search apparatus 11 Information collection part 13 Information storage part 13a Index information 15 Information search part

Claims (5)

[Claims] 1. An information retrieval / management method for presenting a predetermined directory service provider and a category, wherein a similar category is retrieved from a category provided by the directory service by giving an example. Information retrieval and management method. 2. An information search / management method for presenting an optimal directory service provider and a category when a directory service is provided from each of a plurality of providers. Collecting and recording and storing, and when an example is given, the similarity is calculated from the example and the stored category information, and an optimal directory / directory is calculated from the similarity.
2. The method according to claim 1, further comprising the step of presenting a service provider and a category. 3. An information search / management apparatus for presenting a predetermined directory service provider and a category when a directory service is provided from each of a plurality of providers, wherein the category information is collected from the plurality of providers. Information collecting means, information storing means for recording and storing the category information collected by the information collecting means, and when an example is given, the similarity is calculated from the example and the category information stored in the information storing means. An information search / management apparatus comprising an information search means for presenting a predetermined directory service provider and a category based on the similarity. 4. The information search / management apparatus according to claim 3, wherein said information storage means includes a category record table and an information entry record table. 5. The information search / management apparatus according to claim 3, wherein the similarity is an index represented by an inclusion rate and a content rate.