Информационная технология экстракции бизнес знаний из текстового контента интегрированной корпоративной системы

Abstract

Thesis for a candidate degree in technical science, speciality 05.13.06 – Infor-mation Technologies. – National Technical University "Kharkiv Polytechnic Institute". – Kharkiv, 2016. The aim of the thesis is to develop information technology of extraction of business knowledge of integrated corporate system (ICS) based on the information logic models and methods of text content sense processing. The main results are as follows: a logic linguistic model of fact generation from ICS text streams has been developed which is based on surface grammar characteristics of identification of entities of actions and attributes which allows to effectively extract industry specific knowledge about the subjects of monitoring from text content. The thesis further develops the method of comparator identification used for structuring of ICS business knowledge facts relationship. The method allows to classify the attributes of entities according to class relationships due to sense identity of fact triplets which are determined by the comparator objectively. The paper improves the method of determination of actual set of classified entities of a subject domain which is distinguished by an integral use of linguistic, statistical and sense characteristics in the naïve Bayes classifier. The method allows to classify entities extracted according to types determined a priori. The thesis improves the development of information technology of common information space of corporation business activity which allows complicated knowledge generation by means of explicit generalization of information hidden in the collection of partial facts using algebra logic transformations.