Method of support of question-answer function
Anna Lisovenko, Oleh BisikaloThe method which support question and answer system was developed. The offered method allows is automated to carry out linguistically - syntactic processing of input unstructured text information and the user's question for the purpose of generation of the knowledge base of ontologic type. The method builds the answer to the user's question due to preliminary processing of entrance text content in a natural language and accounting of sense of a question. The basis for a method it is knowledge base model with the indistinct relation of sense which applies semantic model of creation of the knowledge base to word forms of the entrance text and a question. Use of komutiativny semi-group of figurative designs as models of the formal theory Th to offers in a natural language is offered. The developed method has expanded interactive opportunities which cover three types, limited behind formal signs and input information, pseudo-dialogue. The method provides a possibility of formation of the answer to a question on the basis of information from several offers of the text that allows to consider semantic value of an interrogative pronoun and syntactic sentence structure in the answer to a question. It is executed an experiment and it is received assessment of the relevance of answers to test questions developed sistemiy in comparison with known by means of methods of expert estimates