Search academic texts

Information × Registration Number 0213U005707, 0111U001965 , R & D reports Title Models and algorithmic tools for developing multiagent systems using ontology-based approach popup.stage_title Head Voronoy Sergey Mikhailovich, Registration Date 10-07-2013 Organization Donetsk National Technical University popup.description2 The report on SRW: 189 p., 41 fig., 23 tab., 101 sources. The object of research: systems of automatic creation of ontologies from texts in a natural language with usage of multiagent approach. The purpose of research: increase of efficiency of methods of selection of semantic links and building of ontologies from semistrucured texts in a natural language. The received results: the method of building of ontologies on the basis of a collection of text documents with usage of preliminary clusterization is developed. The method of detection of semantic relations in the texts, grounded on algorithms of the nearest and mutual the nearest neighbours, is offered and researched. The algorithm of building of ontologies from not structured data is developed, allowing to receive word-combinations of any length. The model of the distributed system of extraction of knowledge and ontologies creation, oriented to usage of the multiagent approach, is developed and researched. Also the technology of information retrieval in semistructured information sources with usage of the knowledge base and metadata is developed. The forecast: the methods and algorithms of automatic construction ontologies developed in this work allow to reduce quantity of the operations executable manually by the expert of a subject domain. The algorithm of detection of semantic relations between concepts can be used as one of the major stages of construction of semantic nets. The technique of reception of terms-word-combinations from sources of not structured data can find the practical application in systems of extract of knowledge, such as automatic construction of dictionaries-thesauruses and development of ontologies of subject domains. Usage of metadata and the knowledge base simplifies process of integration of an intelligence system. Usage of the multiagent approach to extract of knowledge from the distributed informational resources allows to lower load on servers of an intelligence system and to simplify its extension and adapting to new subject domains. Keywords: ONTOLOGY, SEMANTIC LINKS, SEMISTRUCTURED DATA, INFORMATION RETRIEVAL, MULTIAGENT SYSTEM Product Description popup.authors Єгошина Ганна Анатоліївна Бабаков Роман Маркович Вальчук Роман Валерійович Вороной Олексій Сергійович Вороной Сергій Михайлович Патрушев Володимир Олександрович Чубатов Руслан Євгенович Шелєпов Владислав Юрійович Ярош Ірина Вікторівна popup.nrat_date 2020-04-02 Close