Search academic texts

Information × Registration Number 0219U003681, 0117U004726 , R & D reports Title Development of models and methods for identifying knowledge extraction processes in weakly structured information arrays popup.stage_title Head Sharonova Natalia, Registration Date 08-02-2019 Organization National Technical University "Kharkiv Politechnic Institute" popup.description2 The object of the study is the processes of extraction of knowledge in weakly structured text arrays of information. The purpose of the work is the development of models, methods and information technologies for the intellectual processing of data and subject knowledge for the automated identification of extraction processes of knowledge in weakly structured arrays of information. The research methods are based on the integrated use of the methods of the theory of intelligence, the apparatus of the algebra of relations and the algebra of operations on relations, algebra of predicate operations and the algebra of subtractive operations, the method of comparative identification, theoretical foundations and practical methods of computer lexicography, quantitative and corpus linguistics. Methods are developed and information technologies of automatic extraction from the texts of the subject area of ??lexicographic units and named entities are created. A workshop was created to remove the Ukrainian-language entities. A special information-lexicographic support for solving the problem of extracting the named entities of the Ukrainian language has been developed. For experimental research, prototypes of programs for the automatic extraction of lexicographic units and named entities are created. The evaluation of efficiency is carried out separately for two main tasks that are solved in the study: tasks of extracting the terms of the subject domain from the texts and the task of extracting named entities. For the first problem, Recall = 0.89, Precision = 0.94, for the second - Recall = 0.67, Precision = 0.89. Comparison of the results with the results of similar systems showed the effectiveness of the work of information systems in solving the two problems presented, and also confirm the excellent advantages of the results of the proposed models and information technologies in the analysis compared with domestic and foreign analogues. The obtained results can be used in the form of mathematical, algorithmic, informational, software and other support of the system of automated creation of lexicographic resources of various purposes, as well as systems of automated mining of named entities. The results obtained in the study are implemented as packages of applications for the creation of industry information resources, including lexicographic, thematic, patent and information retrieval. Product Description popup.authors Бабкова Надія Вікторівна Борисова Наталя Володимирівна Гулієва Діна Олександрівна Каніщева Ольга Валеріївна Кочуєва Зоя Анатоліївна Купріянов Євген Валерійович Оробінська Олена Олександрівна Петрасова Світлана Валентинівна Хайрова Ніна Феліксівна Шабанова-Кушнаренко Любов Володимирівна Шаронова Наталія Валеріївна popup.nrat_date 2020-04-02 Close