KR100963667B1 - Apparatus of semantic technological intelligence language mining system for large size database - Google Patents

Abstract

The present invention relates to a device for discovering a semantic-based technical term of a large database, and in particular, to search for and output query terms based on technical terms and contextual information of new and seed to search for information of a specific technical field from a management database. Way; TRS means for extracting a document set including a technical term input from the ALM means and a query word based on contextual information, and corresponding posting information from a scientific information database; Analysis means for extracting a technical term and contextual information from a document set and posting information provided by a TLS means and analyzing the relations between the technical terms; Tracking means that receives technical term, context information, association information and document set from analysis means, and traces and extracts technical knowledge based on occurrence frequency, association, and expansion including technical time of occurrence, location, author ; ARS means for receiving technical terms, contextual information, and association information extracted by the analysis means, extracting new technical terms and contextual information, and recording them in a management database; And an ERA means for accessing the ALS means and extracting and providing description terms, context information, relations and description documents from external resources; It is characterized in that the configuration, including, to increase the overall search efficiency and utilization of large databases, and to analyze and accumulate the relationship of the searched technical terms to quickly search and track the relationship, time series analysis, classification, etc. of the technical information in real time This speeds up technology review, development, and decision making.

Patent, Paper, Database, Search, Query, Technical Terms, Contextual Information, Extraction

Description

Meaning-based technology term extraction device for large-scale database {APPARATUS OF SEMANTIC TECHNOLOGICAL INTELLIGENCE LANGUAGE MINING SYSTEM FOR LARGE SIZE DATABASE}

The present invention finds technical terms from a database for recording and managing scientific and patent information. In particular, when a large amount of information is managed in a primitive text database, it takes a lot of time to search in a conventional manner. The present invention relates to a device for discovering a semantic-based technical terminology of a large-scale database which extracts, refines, organizes, links, expands, and analyzes the content of text.

Human beings have accumulated the information of experiences, knowledge, skills, etc. acquired through records, and the next generation's descendants have continued to develop culture and live abundant life by repeating the process of further developing by using accumulated information.

The accumulated amount of this information increases over time, and generally used as a form of accumulation, and systematically manages many books with various knowledge and information, and finds the required knowledge and information quickly and accurately when desired. It is also important to pay.

With the development of computers, various knowledge and information have been recorded and managed in text, and the database is developed to accurately and quickly find the necessary contents among the information managed by computers as described above. It became.

Today, the amount of knowledge and information accumulated in texts is rapidly expanding in all fields including humanities, society, and science, and computers are essential for recording and managing such knowledge and information.

In particular, as the modern society becomes industrialized, the necessity of information retrieval for technology development is increasing, and mining the desired or required technical information from the accumulated large-scale technical information becomes another independent technical field. .

Referring to Figure 1 describes the concept of discovery of general technical information, generally includes technical information, there are patent documents, papers, technical reports, such technical information is recorded and managed in a text-oriented database (DB) do.

As described above, the text information of the database in which the technical information is recorded is searched through software (SOFTWARE: S / W) through analysis (ANALYZE) and explanation (REPRESENTATION).

The technical information retrieved as described above is a technical knowledge (REPERTOIRE) as a technical knowledge (REPERTOIRE), and should be recognized by managers and managers, and supports judgment or decision.

In other words, the information provided is verified by experts, technically analyzed, and used as a reference for decision and judgment by the CEO or manager based on the results of the research. Therefore, the discovery of technical information is a very important means of decision making in modern industrialized society.

However, as the amount of text-based technical information accumulated and managed by a computer increases greatly, there is a problem in that it takes a long time to search, analyze, and utilize desired information.

In addition, the meaning of certain words and terms used by the development of the technology is changed to be used in different meanings or expanded and reduced, used in other similar terms or the same term in different meanings, while at the same time different technologies Already developed or used in.

Therefore, text information managed in large quantities, especially scientific information, is automatically processed in advance, and analyzed and provided by CLARIFICATION, LIAISON, ARRANGEMENT, and EXTENSION technical knowledge. There is a need to develop technologies that increase efficiency.

The present invention has been made to improve the above-mentioned problems and necessities, in particular, by analyzing the text information of the database to extract the technical terms by keywords, iteratively repeating the complex technical terms connected to each technical term and a plurality of technical terms The purpose of this study is to provide a semantic-based technical term extraction device for large-scale databases that improves the efficiency of search because it is extracted and managed by TECHNOLOGICAL INTELLIGENCE.

In addition, the present invention is to provide a device for discovering the meaning-based technical terminology of a large-scale database that analyzes the technical terminology of scientific and technological information, which is managed as a database and includes articles and patents, to classify the relationship between each technical information in real time. Purpose.

The present invention has been made in order to achieve the above object, the ARM means for retrieving and outputting a query based on the technical terms and context information of the new and seed to search for information of a specific technical field from the management database; TRS means for extracting a document set including a technical term input from the ALM means and a query word based on contextual information, and corresponding posting information from a scientific information database; Analysis means for extracting descriptive terminology and contextual information from a set of documents and posting information provided by a TLS means, and analyzing an association of descriptive terms; Tracking means that receives technical term, context information, association information and document set from analysis means, and traces and extracts technical knowledge based on occurrence frequency, association, and expansion including technical time of occurrence, location, author ; ARS means for receiving technical terms, contextual information, and association information extracted by the analysis means, extracting new technical terms and contextual information, and recording them in a management database; And an ERA means for accessing the ALS means and extracting and providing description terms, context information, relations and description documents from external resources; Present a configuration comprising a.

Preferably the DRS means for analyzing the technical knowledge accumulated by the tracking means for a predetermined time to provide a service scenario; A knowledge database for recording and managing new technical knowledge analyzed and tracked by the analysis means and the tracking means; A management database that receives, records, and manages technical terms and context information from an ALS means and provides them to the ALM means and analysis means; And a scientific information database connected to TLS means for recording, managing and providing patents, articles, technical reports, and technical documents extracted from the outside by means of the RS; Presents a configuration that includes more.

In addition, the ALM means extracts a query word that searches a document set of a specific description area and a technology pattern with a relatively high number of selections based on the new and seed description terms inputted from the management database, and inputs the information from the management database. It is characterized by consisting of a configuration of extracting a query word of the frequent vocabulary pattern by the context information of the new and the seed and providing to the TLS means.

Also, the TLS means inputs a query term of a technical term and context information from an ALM means and includes a set of documents including the query word, a designated part-of-speech, and a sentence component from the scientific information database in a state where a load is operated below a predetermined value. And extracting the posting information of each document.

The posting information may include weight information of an index word constituting each document.

The analyzing means may further include tass means for inputting a document set and posting information from the TLS means and extracting a technical term and context information; A Tama means for inputting document set and posting information from a TLS means, inputting a new terminology set from a knowledge database, and extracting technical association information between documents and providing the information to the ALS means and the knowledge database; A TLA means for inputting the results extracted from the TAS means and the Tama means, providing the term candidates to the TAS means using the thesaurus, ontology, and lexical intelligent network, and providing the relations of the technical terms to the Tama means; Characterized in that consisting of a configuration including a.

In addition, the TAS means extracts a technical term from a document set input from a TLS means by pattern analysis of a part-of-speech arrangement method, classifies new technical terms not recorded in the management database in an automatic and manual manner, and classifies the technical terms. Extract the contextual information by the user and provide it to the ALS means, extract the technical terminology used with the queried contextual information, and classify the new technical terminology not recorded in the management database in an automatic and manual manner. It is characterized by consisting of a configuration that extracts the contextual information by the term provided to the ALS.

In addition, the Tama means extracts an association relationship from the parts of the documents input from the TLS means by the part-of-speech information pattern and vocabulary of the technical term, automatically reconfirms the existing association information, and automatically and manually extracts the extracted association relationship. It is characterized in that the configuration consisting of a thesaurus, ontology, vocabulary intelligent network to analyze the association to provide the ALS and the association information by a specific technology to the knowledge database.

In addition, the analysis of the relationship by the thesaurus, ontology, and lexical intelligence network is characterized by the analysis by TLA means.

In addition, the tracking means comprises: a satt means for inputting the technical terms and document content and the relationship information extracted from the analysis means and tracking the technical knowledge manually and automatically; Sam means for inputting descriptive terms, associations, document content information from the Sat means, and inputted descriptive terms, associations, document information information from the knowledge database to provide statistical analysis results; And t-SIM means for inputting description terms, association relations and document contents information from the satt means, and inputting description terms, association relations and document contents accumulated from the knowledge database, and classifying them into description sets and providing clustering results. Characterized in that consisting of a configuration including a.

In addition, the sart means inputs the technical terms and document contents and association information extracted and verified from the analysis means, and inputs the accumulated technical terms, technical and technical information, and linkage information by technical and document from the knowledge database. It is characterized by consisting of a technical knowledge that tracks the associations including occurrence time, producer information, and location information for each technical term to the knowledge database and the DSA means, and provides them with charts and table information.

In addition, the association relationship is characterized in that it includes the fusion and fragmentation of the technology according to the occurrence frequency, relationship, distance information of the technical term, occurrence frequency for each occurrence position, estimation and verification of a new name.

In addition, the sampling means inputs information of technical terms, relations, and document contents from the Sat means and the knowledge database, and hypothesizes the document set of texts provided in large quantities through the Sat means based on the hypothesis setting, the frequency information and the corresponding cause. Nonlinear regression, principal component analysis characterized in that consisting of a configuration for tracking by analyzing any one or more selected from.

In addition, the TSI means input information of the technical terms, associations, document contents from the Sat means and knowledge database, and the relationship between the name and description extracted from the document set of the text provided in a large amount through the Sat means Characterized in that consisting of a configuration to classify and cluster.

The management database may include a technical term dictionary database that accumulates and manages the seed terminology; And a context information database for accumulating and recording seed context information; Characterized in that consisting of a configuration including a.

The scientific information database may further include a patent database that records patent information collected from domestic, US, Japan, and European regions as patent information; An article database that records English abstracts of articles and technical reports managed by thesis information; An external database for collecting and recording technical documents including patents, articles, and technical reports from outside; Characterized in that the configuration consisting of.

In addition, the ALS means records the new technical terms, context information and association information extracted by the analysis means and the technical terms, context information and association information extracted from the outside by the means in the assigned area of the management database. And characterized in that consisting of a configuration to manage.

The EAL means may be configured to automatically extract technical terms, contextual information, and association information from external web sites and databases, and provide them to the knowledge database, the scientific information database, and the AR means. .

The present invention devised in order to achieve the above object, including the means of AM, TS, analysis means, tracking means, management database, scientific information database, AI means, AI means, knowledge database means In the device for discovering the technical terminology, the technical term is extracted automatically and manually by the pattern analysis of the part-of-speech method from the document set and posting information of the specific technology provided by TLS means, and the new technical term and context information are classified. Analysis means for analyzing and verifying the relations between the extracted technical terms; And the technical terms analyzed by the analysis means are extracted and estimated from the document set to indicate the relation of expansion status. Expansion, frequency of occurrence by time, relationship between technical terms, technology fusion and fragmentation by distance information, frequency of occurrence by location Tracking means for estimating and verifying a new name; Present a configuration comprising a.

Preferably, the analysis and verification of the relationship between the extracted technical terms is characterized in that consisting of the configuration using any one selected from thesaurus, ontology, lexical intelligence network.

The present invention devised in order to achieve the above object, an apparatus comprising an ALM means, a TS means, an analysis means, a tracking means, a management database, a scientific information database, an ALS means, an AI means, a knowledge database In the method of discovering the semantic-based technical term of a large database by means of analysis, when the operational load of the scientific information database is low by a predetermined ratio by the analyzing means, the AML means extracts a query searched from the management database and provides it to the TLS means. And the TLS means extracting the document set of the designated technology and the corresponding posting information from the scientific database by the query and providing the analysis information to the analyzing means; The analyzing means may include: a technical term process of notifying an ALS means and registering in a knowledge database when a technical term of a new and seed is detected from a document set by a third means; The analyzing means includes: a context information process of notifying the ALS means of registering new and seed context information from the document set by the Tama means and registering it in the knowledge database; Providing information in the document set and knowledge database to the tracking means to track the associated skills, providing the tracked information to the knowledge database and the means for RS, and analyzing and outputting the tracked information as a document; Present a configuration comprising a.

Preferably, if the operational load of the scientific information database is higher than a predetermined ratio, query means are collected from an external resource by the AL means, notified to the AR means, registered in the knowledge database, and the collected technical documents are stored in the scientific information database. A recording process that proceeds to the outputting process after recording in the; Present a configuration that includes more.

In addition, the tracked information may be related to the expansion and contraction of technical terms, the occurrence frequency analysis by the occurrence time, the analysis of the relationship between the technical terms by the generation time, the fusion and division of the technology by the distance information between the technical terms. Process analysis, frequency of occurrence by location, and estimation and verification of new names.

The present invention having the above configuration has an industrial use effect of increasing the overall search efficiency and utilization of a large database because it continuously and repeatedly extracts and manages technical terms and contextual information based on the meaning of scientific and technological information made of text.

In addition, by analyzing and accumulating the relations of technical terms retrieved from the texts of scientific and technological information, the user-friendly effect of facilitating the technology review and development by quickly searching and tracking the relations, time series analysis, and classification of technical information in real time. have.

Hereinafter, with reference to the accompanying drawings, according to the present invention having the configuration as described above, a preferred embodiment of the semantic technology term extraction apparatus of a large-capacity database will be described in detail.

Example

It is attached to explain the present invention, Figure 2 is an example of the present invention is a diagram illustrating the step-by-step operation to achieve the final goal of the semantic technology term extraction apparatus of the database, Figure 3 is an example of the present invention Figure 4 is a functional configuration diagram of the semantic base technology term extraction device of the database, Figure 4 is a detailed functional configuration diagram of the analysis means constituting the meaning-based technical term excavation apparatus of the database according to an example of the present invention, Figure 5 Fig. 6 is a detailed functional configuration diagram of the tracking means constituting the semantic-based technical term excavation apparatus of the database as an example. 7 is a configuration diagram and FIG. 7 is an example of the present invention. FIG. 8 is a detailed functional configuration diagram of a scientific information database constituting a semantic technology term discovery device, and FIG. 8 is a detailed functional configuration diagram of a knowledge database of analysis means constituting a semantic technology term discovery device of a database according to an example of the present invention. FIG. 9 is an explanatory diagram of a relation of technical terms tracked by an apparatus for discovering meaning-based technical terms of a database according to an example of the present invention, and FIG. 10 is a meaning-based technical term of a database according to an example of the present invention. 11 is an overall detailed functional configuration diagram of the excavation apparatus, and FIG. 11 is a flowchart illustrating a method of discovering a term based on technical terms of a database according to an example of the present invention.

In describing an example of the present invention, drawings and descriptions of well-known technical contents that are not directly related to the present invention are omitted, and thus the present invention is clearly communicated without obscuring the gist of the present invention.

Referring to Figure 2 describes the step-by-step operation to achieve the final goal with a device for discovering the semantic technology terminology of the database by means of an example of the present invention, search for text-based technical information, including patents, articles, technical reports, etc. Search for the technical information you need using the management system.

Technical information retrieved as described above, using the technical vocabulary resources, knowledge expression technology, language processing technology, machine learning technology, etc., scientific and technical vocabulary network, terminology automatic recognition technology, scientific and technical expertise, rule-based information extraction, text-based Automatically detects technologies, extracts time series relationships, discovers relationships between skills, classifies them automatically, and associates them with novelty extraction, topic clustering and settlement, and discovery rule discovery. Recognizes and extracts information and analyzes it in conjunction with external information in real-time.

The technical information processed in real time as described above is personalized based on user profile with reference to externally linked information, revolutionized functionalized based on machine learning, adapted to topic ranking based on user feedback, and personalized secret means. This allows the user to reorder components by user scripts.

That is, by selecting and changing the technical information and contents to be searched in various ways, and analyzing the connection relations based on time, region, and technology, the user supports the decision making easily.

By operating in the real-time dialogue method of the above system, it provides technical knowledge, accurate information search, analysis of search results in real time, using the latest natural language processing method, pattern recognition, data discovery method, etc. It targets a large number of RESOURCEs, including patents, papers, technical reports, and web documents.

Referring to FIGS. 3 to 8, an apparatus for discovering a semantic technology term based on a database according to an example of the present invention will be described. An ARM means 100, a TRS means 110, and an analysis means may be described. 120, tracking means 130, management database 140, scientific information database 150, ARS means 160, ERA means 170, DRS means 180, the configuration includes the knowledge database 190.

The ALM means 100 searches for and outputs a query word based on new and seed technical terms and context information for searching for information of a specific technical field from the management database 140. The management database 140 accumulates and records the technical term dictionary database 142 that accumulates and manages the previously managed seed description terms and the previously managed seed context information as shown in FIG. 6. And a contextual information database 144 that manages and manages.

That is, the ALM means 100 has a relatively large number of selections based on the new and seed technical terms inputted from the technical term dictionary database 142 of the management database 140, and generates a document set of a specific technical area and technical pattern. Extracts a query word that can be searched, extracts a query word of a lexical pattern that frequently appears by new and seed context information input from the context information database 144 of the management database 140, and the extracted technology as described above. The term and contextual information are provided to the TS unit 110.

The technical term is a term commonly used in each technical field, and the meaning included in the term is a word easily understood and understood by a person having ordinary knowledge in the technical field. The new technical term is a newly searched technical term, and the seed technical term is a technical term that is previously searched and recorded and managed in a database (DATABASE: DB).

The query term based on the terminology is a term used to select a term or a term desired by a user or a user from among a plurality of terminologies recorded and managed in the terminology dictionary database 142. For example, “STEM CELL” and “GENE”. EXPRESSION "and the like, and the query term based on the context information is recorded and managed in the context information database 144. In order to extract the relations of the new terminologies, the term is likely to appear in a specific technical name or relationship. As a query word generated by the appearing lexical pattern, it may be, for example, "TECHNOLOGIES SUCH AS", "AND OTHER TECHNOLOGIES", and the like.

The technical idea of the present invention is to extract and analyze a necessary technique for a "document set" that includes a specific technical area or pattern using a query rather than analyzing the entire database to extract the name or relationship of the technology. In other words, it is effective to select and retrieve only information of a specific technical field by using a query from a large database.

The TRS means 110 extracts from the scientific information database 150 a document set including the technical term and the query word based on the context information input from the ALM means 100 and corresponding posting information. The scientific information database 150 records an English abstract of a patent database 152, a paper and a technical report, which records and manages patent documents in Korea, the United States, Japan, and Europe, as shown in FIG. And an external database 156 that collects and records a technical document including a patent database, a patent, a paper, and a technical report from the outside including the web.

The TS unit 110 inputs a query term of a technical term and context information from the ALM unit 100 and the science information database 150 is operated with a load below a predetermined value. The information database 150 extracts the set of documents including the query word, the designated part-of-speech, the sentence component, and the posting information of each document.

The posting information is weight information of index words formed in each technical document.

The TS unit 110 has a large capacity, for example, about 50 gigabytes (GIGA BYTE) to the article database 154 and about 40 gigabytes to the patent database 152 based on the query word generated by the AR means 100. Search scientific information database 150 having a total capacity of 90 gigabytes of bytes. The result of the search is a collection of documents containing a particular query.

The search by the query is to search and extract based on language information by parts of speech and sentence components. For example, the search based on <noun phrase> + "AND OTHER" + "TECHNOLOGIES" includes "AND OTHER TECHNOLOGIES". Searches only documents that have "noun phrase" in front of them, and in another example, searches based on "TECHNOLOGIES" + "ESPECIALLY" + <noun phrase> + <noun phrase> + ,,,, are included by TECHNOLOGIES ESPECIALLY. Only those documents that are listed after the "noun phrase" will be searched.

Further, internal analysis information for each individual document is provided. For example, weight information (DOCUMENT FREQUENCY, TERM FREQUENCY, etc.) of index words is analyzed and provided.

That is, the TS unit 110 inputs a query term based on technical terms and context information from the AL means 100, searches the scientific information database 150, and extracts a document set including the query term. At the same time, since the posting information including the index weight information of each document is extracted together, a document set of a specific technical area is extracted and provided to the analyzing means 120.

The analyzing means 120 extracts the technical terms and the context information from the document set and posting information provided by the TLS means 110, and analyzes the relationship between the technical terms, TAS means 122, The configuration includes a TLA means 124 and a Tama means 126.

The TAS means 122 inputs a document set and posting information from the TLS means 110 and extracts a technical term and context information. The technical term extracts a technical term from the document set by pattern analysis of a part-of-speech arrangement method, and a management database. Automatically and manually classify the new technical terminology not recorded in 140, extract the contextual information by the classified technical term, and provide it to the ALS means 160, and the technology used with the queried contextual information. The terminology is extracted and the new technical terminology not recorded in the management database 140 is automatically and manually classified. The contextual information based on the classified technical term is extracted and provided to the ALS unit 160.

The tass means 122 is a technology acquisition system, and inputs a document set and posting information to extract document descriptions and contextual information including the extracted terminology, the publication year or time of the document, the author, the publication location, and the like. That is, the document set extracted from the TLS means 110 is provided, and two (2) modes (MODE) are extracted based on descriptive terms and contextual information.

The extraction mode based on the descriptive term is a method of extracting an unspecified descriptive term from a simple part-of-speech pattern analysis method from each document including the selected descriptive term. When the extracted technical terms are compared with the seed technical terms recorded in the technical term dictionary 142 of the management database 140 and analyzed as new technical terms not recorded and managed, the automatic and manual methods are verified. If it is found to be technical terms by verification, the relevant context information is collected and extracted, and the verified new technical terms and context information are provided to the ALS means 160 for further recording and management in the management database 140, It is provided to the knowledge database 190 and recorded and managed in the technical knowledge database 192, the verified information is recorded and managed in the verification set database 194 of the knowledge database 190.

The automatic verification of the new technical term is a method of verifying the technical term in connection with external information, and the manual verification is a method of manually analyzing and determining by the expert.

The extraction mode based on the context information automatically extracts terms related to specific context information and is classified as new context information in comparison with the seed context information recorded and managed in the context information database 144 of the management database 140. In the same manner as above, the automatic and manual verification is performed, and the verified new context information is provided to the ALS means 160 to be additionally recorded and managed in the management database 140 and provided to the knowledge database 190 to describe it. Recorded and managed in the knowledge database 192, the verified information is recorded and managed in the verification set database 194 of the knowledge database 190.

The TLA means 124 inputs the results extracted from the TAS means 122 and the Tama means 126, and provides the technical term candidates to the TAS means by using the thesaurus, the ontology, and the lexical intelligent network. Provide the Tama means with the technical term associations.

The thesaurus is a term dictionary that is recorded and managed by a computer for information retrieval, and operates in such a manner as to manage synonyms, antonyms, synonyms, upper words, lower words, related words, etc. of each term.

The ontology generally refers to a study or interest in what kind of reality exists in the universe, and is derived from the compound word 'LOGIA', which means 'ONTO' in Greek meaning 'reality' and 'thesis or lecture'. It refers to a field that originates and seeks research on the essence of things.In computer science and information science, it is a data model that expresses a specific area. It is a set of formal vocabulary that describes the relationship between concepts, and reasoning (REASONING, INFERENCE). It is applied to the semantic web, semantic web service, and the like, which are used to describe relations with Internet resources in a specific field.

The TLA means 124 is a set of language processing modules optimized for a database of science and technology, an in-depth sentence analysis module for extracting and verifying associations of technologies, and considering difficult sentences and various sentence expression methods. SHALLOW PARSER that applies partial syntax analysis technology, and analyzes the ambiguity of machine learning-based part-of-speech learned with specialized information learning documents. It is a part-of-speech tagging system with optimized operation speed to efficiently analyze large amounts of information. .

The TMA means 126 inputs a document set and posting information from the TLS means 110, inputs a new set of technical terms from the knowledge database 190, and extracts technical relationship information between documents. Providing the information to the 160 and the knowledge database 190, extracting the correlation information, automatically revalidating the validity of the previously extracted and managed correlation information, and automatically and manually checking the extracted correlation information. After verifying, the association information is analyzed by the thesaurus, ontology, and lexical intelligence network and provided to the ALS 160, and the association information by the specific technology is provided to the knowledge database 190.

The tama means 126 may be configured to compare the new technical terms accumulated in the knowledge database 190 with the seed technical terms accumulated in the management database 140 from the document set extracted by the TLS means 110 to the present. Correlation is extracted based on lexical and part-of-speech information patterns. For example, extraction of technical terms based on prepositions is the same as “protein kinase IN heat treat protein”, and the extraction of technical terms based on verb phrases is called “SURFACE WATER” ARE ACCOMPANIED. BY HIGH FLUXES OF ORGANIC MATTER, "and others, such as" CHRONOLOGICAL AGE THAT IS APPLICABLE TO THE NATURAL POPULATION. "

In addition, since the tama means 126 re-verifies previously extracted correlation information, the CLARIFICATION operation is automatically performed, and the extracted correlation information is automatically and manually verified as described above. In addition, the relationship analysis using thesaurus and lexical network (eg WORDNET) is classified into two types. For example, the association information in the form of GENERIC RELATION is the same as the “parent”, “child”, “effect”, “cause”, etc., and the association information in the SPECIFIC RELATION form can be derived between a specific area or technical term. Is a special form of relationship, such as “EIGEN-VALUE PROBLEM OF AN OPEN SYSTEM UNDER STRONG INTERACTION”.

The tracking means 130 receives the technical term, the contextual information, the association information and the document set from the analysis means 120, and the relation of occurrence frequency, association, and extension including the occurrence time, location, and author of the technical term. By extracting the technical knowledge by means of the Sam (SAM) means 132, Sat (SATT) means 134, TSI (TCM) means 136 is a configuration that includes.

The sart means 134 inputs the technical terms, document contents, and association information extracted from the TAS means 122 and the Tama means 126 of the analysis means 120, and accumulates from the knowledge database 190. The technical knowledge is traced manually and automatically by inputting new technical terminology, technical and technical information, and technical and document linkage information. Each of the traced technical terms is technical knowledge of an association relationship including time of occurrence, producer information, and location information, and is provided to the knowledge database 190 to be recorded and managed, and provided to the DLS means 180 to provide various services. Applied and printed as a document with table and table information.

The association includes the fusion and fragmentation of the technology according to the frequency of occurrence of the term, relationship, distance information on the database, the frequency of occurrence by position, and the estimation and verification of a new name.

The sat means 134 inputs the correlation between the extracted and verified technical terms and document content information such as the year or time of publication of the document, the author, the publication location, and the technical terms, and records the traced technical knowledge into the knowledge database 190. ) To record management, to provide to the RS means 180 to be utilized in various application services, and output as technical document information by charts and tables. That is, all the extracted technical documents include time information by the year of publication, producer information by the author, and location information by the generation place, and the like, and the relationship can be tracked in various forms using the document information.

As an example of the above tracking, the association can be traced in the form of [technical term]: [time of occurrence]: [author]: [location of occurrence], and also [technical term]: [association relationship]: [technical term]. Relationships can be tracked in the form: term: [time of occurrence]: [author]: [location of occurrence].

The relationship between each of the technical documents tracked as described above may be expressed as shown in FIG. 9.

In FIG. 9, each point is an extracted technical term, and a line connected between each technical term is an extracted relation.

The correlation extracted as described above may be estimated by the concept of EXTENSION. For example, if we trace the relationship between technical term A, technical term B, and technical term C, technical term A is related to technical term C because A-> B, B-> C eventually becomes A-> C. It is tracked as having a relationship. In addition, it is possible to track the correlation of the occurrence frequency according to the occurrence time of the technical term, to track the correlation between the occurrence time, and to track the process of fusion or division of the corresponding technology by the distance information operation in the database. Can be analyzed. In addition, the frequency of occurrence of the technology based on the location information can be tracked and analyzed, and the name change of the technical term can be estimated or verified.

The sampling means 132 inputs the technical terms, the relations, and the contents of the documents from the sat means 134, and inputs the technical terms, the relations, the contents of the contents of the documents accumulated from the knowledge database 190, and statistical analysis results. At least one selected from the group including hypothesis setting, frequency information and nonlinear regression analysis based on the cause, and PRIINCIPAL COMPONENT ANALYSIS (PCA) analysis. The correlation is tracked by statistical analysis.

That is, the sampling means 132 is a backend module for collecting various statistical information, and receives large amount of numerical information or text information and statistically analyzes it according to the situation.

The TSI means 136 inputs description terms, association relations, and document contents from the satt means 134, and inputs description terms, association relations, and document contents accumulated from the knowledge database 190, and sets the descriptions. By classifying and clustering and tracking the results provided, the document set of the text provided in a large amount through the means 134 is classified (CLASSIFICATION) and clustered (CLUSTERING) by the relationship between the extracted name and technology.

That is, the TSI module 136 clusters and classifies technical documents extracted using various models, for example, NAIVE BAYESIAN MODEL, k-NEAREST NEIGHBOR MODEL, SVM MODEL, etc. Classify.

Referring to FIG. 4, the analyzing means 120 includes a TAS means 122, a TLA means 124, and a Tama means 126.

Referring to FIG. 5, the tracking means 130 includes a leakage means 132, a sat means 134, and a TSI means 136.

Referring to FIG. 6, the management database 140 will be described by accumulating and recording the seed terminology provided by the ALS means 160 and accumulating and recording the technical term dictionary database 142 and seed context information. And a contextual information database 144 that manages and manages.

Referring to FIG. 7, the scientific information database 150 will be described. A patent database 152 for recording and managing patent information collected from Korea, the United States, Japan and Europe will be recorded. It is a configuration that includes an external database 156 that collects and records a technical document including a paper database 154, a patent, a paper, and a technical report from the outside and the web (WEB).

Referring to FIG. 8, the knowledge database 190 is a configuration including a technical knowledge database 192 and a verification set database 194, and FIG. 10 is an entire functional block of a device for discovering a semantic-based technical term of a database according to the present invention. This is shown in detail.

Hereinafter, referring to the attached FIG. 11, a method of discovering a semantic-based technical term of a database according to an example of the present disclosure will be provided. Terminology courses; Contextual information process; Output process; Recording process; It includes a configuration.

A database using the configuration as described above, which includes an ALM means, a TS means, an analysis means, a tracking means, a management database, a scientific information database, an ALS means, an ALA means, a DLA means, and a knowledge database. This section describes in detail how to discover semantic-based technical terms.

It is determined by the analyzing means whether the operating load of the scientific information database is low at a predetermined rate (S100). That is, the method according to the present invention is automatically and repeatedly operated when the operating load is small by determining whether the user or the user's connection is large or small.

When the operation ratio of the scientific information database is low at a predetermined ratio in the determination (S100), the AL means provides a query word extracted by searching the management database to the TS means (S110). The query word is a query word based on descriptive terms and contextual information.

The TLS means extracts a document set and posting information from the scientific information database by using a query word and provides it to the analyzing means (S120). That is, a set of documents classified into a specific technical field and the query and posting information are searched and extracted from a scientific information database, and provided to analysis means, and the posting information includes weight information of index words constituting a document.

The analyzing means checks whether the technical terms of the new and seed are detected or extracted from the provided document set (S130), and if the technical terms are detected or extracted, notifies the ALS means to the knowledge database. Register (S140). That is, the ALS means registers in a technical term dictionary database constituting a management database, and a knowledge database is registered in a corresponding technical knowledge database.

In addition, the analyzing means checks whether the context information of the new and the seed is detected or extracted from the provided document set (S150), and if the technical term is detected or extracted, notifies the ALS means to the knowledge database. Provide and register (S160). That is, the ALS means registers in the context information database constituting the management database, and the knowledge database registers in the corresponding technical knowledge database.

The analyzing means provides the tracking means with the document set extracted by the query word, and the knowledge database records and manages and provides all the new knowledge information including the technical term to the tracking means (S170).

The tracking means receives the corresponding information including the document set and the technical term, and tracks whether there is an associated technology and whether there is an associated relationship including the technical term, the occurrence time, the producer information, and the location information.

Check whether there is an associated technology by the above tracking (S180), and if it is determined that there is an associated technology, the tracked technical information is provided to a knowledge database and a DSA means, recorded and applied, and outputted as a document by charts and tables. (S190).

In addition, when the operational load of the scientific information database is high at a predetermined rate, query words, technical literature information, and the like are collected from external resources including the web by EAL means and provided to the external database of the RS means and the scientific information database. Record (S200), and proceeds to the output process (S170).

The present invention having the above configuration has the advantage of searching and extracting a large amount of information in real time because of limited search and extraction of technical documents in a specific field such as natural language processing, pattern recognition, data discovery, and the like.

Although the present invention has been described as an example, it is not necessarily limited to such an example, and various modifications can be made without departing from the technical spirit of the present invention. Therefore, the examples disclosed in the present invention are not intended to limit the technical idea of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by the examples. The scope of protection of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

1 is a conceptual diagram of excavation of general technical information;

FIG. 2 is an explanatory diagram of step-by-step operation for achieving a final goal with an apparatus for discovering a term based on technical terms according to an example of the present invention;

3 is a functional configuration diagram of a semantic technology term extraction means based on the database according to an example of the present invention,

4 is a detailed functional configuration diagram of an analysis means constituting a device for discovering a semantic based technical term of a database according to an example of the present invention;

5 is a detailed functional configuration diagram of a tracking means constituting a device for discovering a term based on technical terms by means of an example of the present invention;

6 is a detailed functional configuration diagram of a management database constituting a semantic-based technical term discovery device of a database according to an example of the present invention;

7 is a detailed functional configuration diagram of a scientific information database constituting a device for discovering a term based on technical terms by an example of the present invention;

8 is a detailed functional configuration diagram of a knowledge database of analysis means constituting a semantic-based technical term discovery device of a database according to an example of the present invention;

FIG. 9 is an explanatory diagram of association of technical terms tracked by an apparatus for discovering a term based on technical terms by an example of the present invention; FIG.

10 is an overall detailed functional configuration diagram of a device for discovering semantic based technical terms of a database according to an example of the present invention;

11 is a flowchart illustrating a method of discovering a term based on technical terms of a database by an example of the present invention.

          ** Explanation of symbols on the main parts of the drawing **

100: ARM means 110: TRS means

120: analysis means 122: TAS means

124: TLA means 126: TAMA means

130: tracking means 132: SAM means

134: SATT means 136: TCM means

140: management database 142: technical glossary DB

144: contextual information database 150: scientific information database

152: Patent Database 154: Thesis Database

156: external database 160: ARES means

170: ERA means 180: DRS means

190: Knowledge Database 192: Technical Knowledge Database

194: Validation Set Database

Claims (8)

ARM means for searching for and outputting query terms based on technical terms and context information of new and seed to search for information of a specific technical field from a management database; TRS means for extracting a document set including a technical term input from the ALM means and a query word based on context information and corresponding posting information from a scientific information database; Analysis means for extracting description terms and context information from the document set and posting information provided by the TLS means, and analyzing the relationship between the description terms; ARS means for receiving technical terms, context information, and association information extracted by the analysis means, extracting new technical terms and context information, and recording them in a management database; And ERA means for accessing the ALS means and extracting and providing descriptive terms, context information, relations and description documents from external resources; Meaning-based technical term excavation device of a database comprising a. The method of claim 1, The technical term, contextual information, association information, and document set are provided from the analysis means, and tracking and extracting the technical knowledge based on the occurrence frequency, association, and extension relation including the occurrence time, location, and author of the technical term Way; DRS means for analyzing the technical knowledge accumulated by the tracking means for a predetermined time to provide a service scenario; A knowledge database for recording and managing new technical knowledge analyzed and tracked by said analysis means and said tracking means; The management database that receives, records and manages technical terms and context information from the ALS means and provides the ALM means and the analysis means to the ALM means; And The scientific information database connected to the TS unit and recording, managing, and providing a patent, a paper, a technical report, and a technical document extracted from the outside by the RS unit by searching; Meaning-based technical term extraction device of the database further comprising. The method of claim 1, wherein the ALM means, Based on the terminology of the new and seed input from the management database, a query word for a relatively large number of selections and a search for a document set of a specific technical area and technical pattern is extracted, and the context information of the new and seed input from the management database is extracted. Meaning-based technical term extraction apparatus of a database, characterized in that consisting of the configuration to provide a query means of frequently appearing vocabulary pattern by the TLS means. The method of claim 1, wherein the TS means, A set of documents including the query word, the designated part-of-speech, and the sentence component from the scientific information database in which the query term of the technical term and the context information is input from the ALM means and the load is operated below a predetermined value, and the posting information of each document Meaning-based technical term excavation device of a database, characterized in that consisting of a configuration for extracting. The method of claim 2, wherein the analysis means, Tass means for inputting document set and posting information from the TLS means and extracting technical terms and context information; A tama means for inputting document set and posting information from the TLS means, inputting a new description term set from the knowledge database, and extracting the technical association information between the documents and providing them to the ALS means and the knowledge database; A TI-A means for inputting the results extracted from the TAS means and the TAM means, providing candidate candidates with technical terms to the TAS means using a thesaurus, ontology, and lexical intelligent network, and providing the relations of the technical terms to the Tama means; Meaning-based technical term excavation apparatus of the database, characterized in that consisting of a configuration. The method of claim 5, wherein the tass means, Extract technical terms from the document set inputted from the TLS means by pattern analysis in a part-of-speech arrangement method, classify new technical terms not recorded in the management database in an automatic and manual manner, and contextual information by the classified technical terms. Extract and provide to the ALS means, extract technical terms used with the queried context information, classify new technical terms not recorded in the management database in an automatic and manual manner, and use the classified technical terms. Apparatus for discovering the meaning-based technical terminology of a database, characterized in that the context information is extracted and provided to the ALS means. The method of claim 5, wherein the tama means, Extract the association from the document set input from the TLS means by the part-of-speech information pattern and vocabulary of the technical term, automatically reconfirm the existing association information, automatically and manually verify the extracted association, and thesaurus And an ontology and a lexical intelligent network to analyze the association and provide it to the AI, and provide the association information by a specific technology to the knowledge database. The method of claim 2, wherein the tracking means, Satt means for inputting descriptive terms extracted from the analysis means, document content and association information, and tracking technical knowledge manually and automatically; Sampling means for inputting description terms, association relations and document contents information from the satt means, and inputting description terms, association relations and document contents accumulated from the knowledge database to provide statistical analysis results; And TSI M means for inputting information of technical terms, relations and document contents from the Sat means, and inputting information of technical terms, relations and document contents accumulated from the knowledge database, and classifying them into description sets and providing clustered results. ; Meaning-based technical term excavation apparatus of the database, characterized in that consisting of a configuration.

Images