WO2012046904A1

WO2012046904A1 - Device and method for providing multi -resource based search information

Info

Publication number: WO2012046904A1
Application number: PCT/KR2010/007239
Authority: WO
Inventors: 정한민; 김평; 이승우; 이미경; 서동민; 성원경
Original assignee: 한국과학기술정보연구원
Priority date: 2010-10-07
Filing date: 2010-10-21
Publication date: 2012-04-12
Also published as: KR101055363B1

Abstract

The present invention relates to a device and method for providing search information based on multiple resources. The invention comprises a database having stored information on a named-entity dictionary and context patterns; an entity recognition and relationship generation module for recognizing, by resource, entities in an input document using the named-entity dictionary and for generating relationships between the recognized entities using the context pattern information; a knowledge base construction module for mapping and storing, as integrated ontologies, the recognized entities and relationships between the entities from multiple resources which are different from one another, and for expanding an integrated knowledge base by performing reasoning over the integrated ontologies; a search module for obtaining search information by resources corresponding to a search query by searching the integrated knowledge base when a user makes a search query entry; and a search result providing module for providing search results information by weighting the obtained search information according to the percentage of each of the resources, and calculating the obtained results and then sorting the calculated results. Therefore, according to the present invention, search information from a combination of multiple resources is provided based on the term entered by a user, and an information service is available from a variety of perspectives depending on a user's requirements, and search results are provided which have been appropriately weighed in proportion to the amount of each resource inputted.

Description

Apparatus and method for providing multiple resource-based search information

The present invention relates to an apparatus and method for providing multi-resource-based search information, and more particularly, to an apparatus and method for providing multi-resource based search information by recognizing an entity of an input document using an entity name dictionary for each resource, Mapping the relationship between the recognized entity and the entity to multiple integrated ontologies in different multiple resources and performing inference on the integrated ontology to expand the integrated knowledge base, Searching the integrated knowledge base to obtain search information corresponding to the search word for each resource, assigning a weight according to a ratio of each resource to the obtained search information, and calculating search result information obtained by sorting the calculated result Based search information providing apparatus and method.

In general, a search information providing device is a device that collects / classifies / accumulates information and extracts it as needed.

One of the indispensable acts in human intellectual activities in the information society is the collection, accumulation, retrieval, and utilization of information that relies on advanced science and technology. Personally, there is a need to find ways to obtain necessary information and data more quickly and accurately for research and development and decision-making, and to be able to meet diverse information needs in society at large. Therefore, it is important how to construct, maintain and operate a search information providing device that can be satisfied by information search.

Therefore, the search information providing device is a device that collects, processes and processes information and data predicted to be necessary by the information demander, and quickly finds information suitable for the request from the database which is stored in a form that is easy to find and provides to the information requester . In this form, the search result output from the search information providing device may be informed by telephone or transmitted by facsimile, and the search result may be transmitted to a home or office of the requester having the computer via the Internet.

And a reference search / fact search / full-text search according to the kind of information accumulated in the database. A reference search is a search for bibliographic references in the literature that are dealing with the subject that the requester wants to know. In fact, a search can retrieve general, numeric, and factual data. The specialized search is to retrieve not only the bibliographic data but also the entire sentence or the original text from the database which has accumulated the full text of the document, if necessary.

However, the conventional search information providing apparatus has been limited in that it can not present the individual resources such as papers, research reports, and patents independently and merely presenting the respective information related to science and technology.

Therefore, it is possible to provide fusion information according to technical fields and to provide the fusion information according to the technical fields, which is necessary for the R & D and research trends of the technical field, by providing the search information that fuses different multi- And to make use of science and technology.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems of the related art, and an object of the present invention is to provide search information fused with multiple resources to a search word input by a user, Based search information providing apparatus and method capable of providing a search result obtained by calculating a weight according to a ratio of resources.

It is another object of the present invention to provide a multi-resource-based search information providing apparatus and method that fuses different multi-resources to enable information service in various viewpoints according to a user's request.

According to an aspect of the present invention, there is provided a database management method comprising: storing a database storing entity name dictionary and context pattern information; recognizing an entity of an input document using the entity name dictionary for each resource; And a knowledge base for expanding the integrated knowledge base by performing inference on the integrated ontology and storing the relation between the recognized entity and the entity in a different integrated resource A search module for searching the integrated knowledge base for search information corresponding to the search word when the search word is input by a user, a module, and a user, and assigning a weight according to a ratio of each resource to the obtained search information, And provides search result information obtained by sorting the calculated results, Based search information providing apparatus is provided.

The database includes at least one category, a name of an entity belonging to each category, an entity name to which an identifier of each entity name is mapped, an entity name database storing a dictionary, a context pattern information database storing pattern information about contexts between entities, And a normalization dictionary database.

The object recognition and relationship generation module recognizes a term by performing morphological analysis, syntax analysis, and semantic analysis on the input document, and when the recognized term is registered in the entity name dictionary, Extracting a context between the recognized entities based on the context pattern information and normalizing the extracted context based on the normalization dictionary database when the entity recognizing unit recognizes two or more entities, And a relationship generation unit for generating a relationship between the two.

The entity recognizer assigns a category and an identifier to each recognized entity.

The relation generation unit expresses a relationship between the extracted entity and the generated entity in terms of a subject, a relation name (Predicate), and an object (Object).

The knowledge base building module stores the relationships among the objects created in the different multi-resources or the meta data of the respective resources as an integrated ontology expressing connection relations between multiple resources.

The search result providing module classifies the search information acquired for each resource by research subjects, calculates a weight by a ratio of each resource input by the user to the number of search for each resource per research subject, The results are ranked by the highest research subjects.

In addition, the search result providing module arranges and provides the calculated results in a predetermined order or in a sequence requested by the user.

According to another aspect of the present invention, there is provided a method for providing search information based on multiple resources, the method comprising the steps of: (a) recognizing an entity of an input document by using an entity name dictionary for each resource; (B) mapping the recognized entity to the integrated ontology in different multi-resources and storing the same, and performing inference on the integrated ontology, (C) if the search word is input by the user, searching the integrated knowledge base to obtain search information corresponding to the search word for each resource; (d) And providing search result information in which the calculated results are sorted. The multi-resource-based search This method is provided for information.

Wherein the step (a) recognizes a term by performing morphological analysis, syntax analysis, and semantic analysis on the input document, and recognizing the term as an entity when the recognized term is registered in the entity name dictionary, Extracting a context between the recognized entities based on the context pattern information when the entities are two or more, and generating a relationship between the entities by normalizing the extracted context based on the previously stored normalization dictionary .

Wherein the step (c) includes the steps of providing a search word input screen when a search is requested by the user, receiving a search word and a ratio of a resource and a resource for the search word through the search word input screen, And acquiring search information corresponding to the input search word for each resource.

The step (d) includes the steps of classifying search information obtained for each resource by a research subject, applying a weight according to a ratio of each resource to the number of searches for each resource for each research subject, And ranking them in the order of higher research subjects.

The search result information is provided in a predefined or ordered order requested by the user.

According to another aspect of the present invention, there is provided a recording medium in which a program for performing a method for providing multiple resource-based search information is recorded and readable in an electronic device.

According to the present invention, it is possible to prevent a decision error caused by fragmentary knowledge and to provide an information service in various viewpoints by fusing different multiple resources by providing various services for supporting decision making, There is a possible effect.

In addition, search information obtained by fusing multiple resources with respect to a search word input by a user and providing a search result that calculates a weight according to the ratio of resources when the ratio of each resource is input can be provided. Therefore, it is possible to provide information services from various perspectives.

1 is a block diagram schematically illustrating a configuration of a multi-resource-based search information providing apparatus according to an embodiment of the present invention;

2 is a flowchart illustrating a method of providing search information based on multiple resources according to an embodiment of the present invention.

3 is an exemplary view of a search word input screen according to the present invention;

4 is an exemplary view showing a search result information providing screen according to the present invention;

The foregoing and other objects, features, and advantages of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which: FIG.

1 is a block diagram schematically illustrating a configuration of a multi-resource-based search information providing apparatus according to an embodiment of the present invention.

Referring to FIG. 1, a multi-resource-based search information providing apparatus 100 includes a database 110, an object recognition and relationship generating module 120, a knowledge base building module 130, A search module 150, and a search result providing module 160.

The database 110 includes an object name dictionary database 112 in which one or more categories, object names belonging to each category, object names to which the identifiers of individual object names are mapped, database 112, context pattern information A database 114, and a normalization dictionary database 116 for normalizing the context.

The entity recognition and relation generation module 120 recognizes an entity of an input document by using the entity name dictionary for each resource and generates a relationship between the recognized entities using the context pattern information.

The entity recognition and relationship generation module 120 includes an entity recognition unit 122 and a relationship generation unit 124.

The entity recognizer 122 recognizes a term by performing morphological analysis, syntax analysis, and semantic analysis on the input document, and recognizes the term as an entity when the recognized term is registered in the entity name dictionary. At this time, the entity recognition unit 122 assigns a category and an identifier to each recognized entity.

When the two or more entities are recognized by the entity recognizing unit 122, the relation generating unit 124 extracts a context between the recognized entities based on the context pattern information, And the extracted context is normalized to generate the relationship between the entities.

For example, in the case of the "notebook consists of monitor, SSD, and mobile processor" statement, the entity recognizer 122 recognizes a notebook, a monitor, an SSD, and a mobile processor entity in the sentence.

At this time, since two or more entities have been recognized, the relationship generation unit 124 extracts only the context corresponding to the previously stored context pattern information, for example, " Bebe verb + past participle + preposition ".

Then, based on the normalization dictionary, the relation generation unit 124 unifies the tense of the corresponding context, for example, "was" to "be" and uses similar expressions such as "similar to" and "similar with" and "similar to" to create a relationship between entities by performing context normalization.

The relationship generation unit 124 expresses the relationship between the extracted entity and the generated entity in terms of a subject, a relationship name (Predicate), and an object (Object).

The method for the object recognition and relation generation module 120 to generate the object recognition and the relation is performed for each resource, and the resource refers to a patent, a thesis, a report, an academic document, a standard, a statistic, an electronic document and the like.

The knowledge base building module 130 maps and stores the relationships between entities and entities recognized from different multiple resources into an integrated ontology and extends the integrated knowledge base by performing inference on the integrated ontology.

At this time, the knowledge base building module 130 stores the relationships among the objects created in the different multi-resources or the meta data of the respective resources in the integrated ontology represented by the connection relation among multiple resources.

The integrated knowledge base generated by the knowledge base building module 130 is stored in the knowledge base DB 140.

When the search word is input by the user, the search module 150 searches the integrated knowledge base and acquires search information corresponding to the search word for each resource.

For example, when the user inputs a search term " smart card ", the search module 150 searches the integrated knowledge base to search information about a smart card, such as a patent, a paper, a report, Statistics, and electronic documents.

The search result providing module 160 provides search information obtained by sorting the calculated results by assigning weights according to ratios of the resources to the search information obtained by the search module 150. [

That is, the search result providing module 160 classifies search information obtained for each resource by research subjects, applies a weight according to the ratio of each resource input by the user to the number of search for each resource per research subject, , And the calculated results are ranked and provided in order of higher research subjects.

At this time, the search result providing module 160 may arrange the calculated results in a predetermined order or in a sequence requested by the user.

For example, a case where a user inputs a "smart card" as a search word and inputs a ratio of each resource to 50% of a thesis, 30% of a patent, and 20% of a report.

The search information obtained for each resource was classified into 2 subjects, 3 patents, 5 reports, 5 patents, 5 patents, 1 report, and 3 patents. Ten patents, and one report, the search result providing module calculates the number of searches by multiplying the corresponding resource ratio.

In the case of Company A, 2 * 0.5 + 3 * 0.3 + 5 * 0.2 = 2.9 and Company B 5 * 0.5 + 5 * 0.3 + 1 * 0.2 = 1 * 0.2 = 4.5.

Then, the search result providing module 160 provides search results in the order of the C company 4.5, the B company 4.2, and the A company 2.9 by ranking the calculation results for each research subject.

The search information providing apparatus 100 configured as described above extracts the implicit information through text mining in multiple resources such as a paper, a patent, and a report, merges with the explicit information such as metadata, This is put on the service platform based on Semantic Web technology, enabling information service from various viewpoints.

FIG. 2 is a flowchart illustrating a method for providing search information based on multiple resources by a search information providing apparatus according to an embodiment of the present invention. FIG. 3 is an exemplary view of a search word input screen according to the present invention. FIG. 5 is a diagram illustrating an example of a search result information providing screen according to the embodiment of FIG.

Referring to FIG. 2, the search information providing apparatus recognizes an entity of an input document using an entity name dictionary for each resource (S200), and creates a relationship between the recognized entities using the context pattern information (S202).

That is, the search information providing apparatus recognizes a term by performing morphological analysis, syntax analysis, and semantic analysis on the input document, and recognizes the term as an entity when the recognized term is registered in the entity name dictionary.

The retrieval information providing device then extracts the context between the recognized entities based on the context pattern information and normalizes the extracted context based on the previously stored normalization dictionary when the recognized entity is two or more To create relationships between entities.

After the execution of S202, the search information providing device maps and stores the relationship between the entity and the entity recognized in the different multi-resources into the integrated ontology (S206), and performs inference on the integrated ontology to expand the integrated knowledge base (S208). That is, the search information providing apparatus extracts internal information through text mining in multiple resources such as a paper, a patent, and a report, merges with external information such as metadata, and puts it on a service platform based on a semantic web technology, To provide information services in the Internet.

After the execution of step S208, if the search word is input by the user (S210), the search information providing apparatus searches the integrated knowledge base and acquires search information corresponding to the search word for each resource (S212).

That is, when a search is requested by the user, the search information providing apparatus provides a search word input screen as shown in FIG. Referring to FIG. 3, the search word input screen 300 includes a search word input area 310 for inputting a search word, a resource selection area 320 for selecting a resource and a ratio for the resource, do.

The user inputs a search word desired to be searched in the search word input area 310 and selects at least one of a patent, a thesis, a report, an academic document, a standard, statistics, and an electronic document in the resource selection area 320 . At this time, the user can input a ratio of the selected resources together.

In this case, the resource selection area 320 is configured as a slide bar type in which two resources can be selected. However, it is natural that the resource selection area 320 can be formed in various forms that can select a plurality of resources such as three or four.

In addition, the search term input screen 300 may further include a year range designation area for designating a year range.

When the user inputs a search word and a ratio of a resource and a resource for the search word through the search word input screen 300, the search information providing apparatus searches the integrated knowledge base for search information corresponding to the input search word Acquired by each resource.

After the execution of step S212, the search information providing apparatus provides a weight according to the ratio of each resource to the acquired search information (S214), and provides search result information obtained by sorting the calculated result (S216) .

That is, the search information providing apparatus classifies the search information obtained for each resource by research subjects, and calculates the search information by applying a weight according to the ratio of each resource to the search number of each resource by the research subject. Then, the search information providing apparatus ranks and provides the computed results in the order of higher research subjects. At this time, the search information providing device arranges and provides the calculated results in a predetermined order or order requested by the user.

The search result information providing screen provided by the search information providing apparatus is as shown in FIG. 4, the search result information providing screen 400 includes a search word providing region 410, a resource providing region 420, and a search result providing region 430 do.

The search word providing region 410 displays a search word input by a user, and the resource providing region 420 displays a resource selected by the user and a ratio of the resource.

The search result providing region 430 is a region in which a result obtained by applying a weight according to the resource ratio to the number of searches of each resource by the research subject is provided. In this case, to be.

For example, in the search result providing region 430, the search results are provided in the order of the A company 10, the B company 8, and the C company 7 in the order of higher research subjects.

The user may change the resource or resource ratio using the slide bar of the resource providing area 420. Then, the search result providing area 430 is provided with a search result reflecting the changed resource or resource ratio.

Thus, those skilled in the art will appreciate that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the embodiments described above are to be considered in all respects only as illustrative and not restrictive. The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

As described above, the apparatus and method for providing multi-resource-based search information according to the present invention provide search information that fuses multiple resources to a search word input by a user, and when the ratio of each resource is input, It is suitable for providing the calculated search result and having a high necessity of providing information service from various viewpoints according to the demand of the user.

Claims

A database storing object name dictionary and context pattern information;

An object recognition and relation generation module that recognizes an entity of an input document using the entity name dictionary for each resource and generates a relationship between the recognized entities using the context pattern information;

A knowledge base building module for mapping and storing relationships between entities recognized by different multiple resources into an integrated ontology and extending the integrated knowledge base by performing inference on the integrated ontology;

A search module for searching the integrated knowledge base and obtaining search information corresponding to the search word for each resource when a search word is input by a user; And

A search result providing module for assigning weights according to ratios of the resources to the obtained search information, and providing search result information obtained by sorting the calculated results;

Based search information providing apparatus.
The method according to claim 1,

The database includes:

One or more categories, an object name belonging to each category, an object name to which an identifier of each object name is mapped, an object name in which the dictionary is stored, a dictionary database;

A context pattern information database storing pattern information on a context between entities; And

And a normalization dictionary database for normalizing the context.
3. The method according to claim 1 or 2,

Wherein the object recognition and relationship generation module comprises:

An entity recognizing unit recognizing the term by performing morphological analysis, syntax analysis, and semantic analysis on the input document, and recognizing the term as an entity when the recognized term is registered in the entity name dictionary; And

Extracting a context between the recognized entities based on the context pattern information and normalizing the extracted context based on the normalization dictionary database when two or more entities are recognized by the entity recognition unit, And a relation generator for generating a relation based on the search result.
The method of claim 3,

Wherein the entity recognition unit assigns a category and an identifier to each of the recognized entities.
The method of claim 3,

Wherein the relation generator represents a relationship between the extracted entity and the generated entity as a subject, a relation name, and an object.
The method according to claim 1,

Wherein the knowledge base building module stores the relationship between the objects created in the different multiple resources or the meta data of each resource as a combined ontology expressed by a connection relation between multiple resources.
The method according to claim 1,

The search result providing module classifies the search information acquired for each resource by research subjects, calculates a weight by a ratio of each resource input by the user to the number of search for each resource per research subject, Wherein the search results are ranked in order of higher research subjects.
The method according to claim 1,

Wherein the search result providing module arranges the calculated results in a predetermined order or in a sequence requested by the user, and provides the sorted results.
A method in which a search information providing apparatus provides search information based on multiple resources,

(a) recognizing an entity of an input document by using an entity name dictionary for each resource, and creating a relationship between the recognized entities using context pattern information;

(b) mapping and storing the relationships between entities and entities recognized by different multi-resources into an integrated ontology and extending the integrated knowledge base by inferring the integrated ontology;

(c) when the user inputs a search word, searching the integrated knowledge base to obtain search information corresponding to the search word for each resource; And

(d) assigning weights according to ratios of the resources to the obtained search information, and providing search result information obtained by sorting the calculated results;

Based search information providing method.
10. The method of claim 9,

The step (a)

Recognizing a term by performing morphological analysis, syntax analysis, and semantic analysis on the input document, and recognizing the term as an entity when the recognized term is registered in the entity name dictionary;

Extracting a context between the recognized entities based on the context pattern information when the recognized entities are two or more; And

And generating a relationship between entities by normalizing the extracted context based on a previously stored normalization dictionary.
10. The method of claim 9,

The step (c)

Providing a search word input screen when a search is requested by the user;

Receiving a search word and a ratio of resources and resources for the search word from the search word input screen; And

Searching for the integrated knowledge base and acquiring search information corresponding to the input search word for each resource.
10. The method of claim 9,

The step (d)

Classifying search information acquired for each resource by research subjects;

Applying a weight according to a ratio of each resource to the number of searches of each resource by the research subject;

And ranking the calculated results in order of higher research subjects to provide search information based on multiple viewpoints.
10. The method of claim 9,

Wherein the search result information is arranged in a predetermined order or in a sequence requested by a user.
The recording medium according to any one of claims 9 to 13, wherein the program for performing the multi-viewpoint-based search information providing method is recorded and readable in an electronic device.