DEVELOPMENT OF A SYSTEM OF QUESTIONS AND ANSWERS FOR THE KAZAKH LANGUAGE BASED ON RULE-BASED AND HMM

Abstract

This study describes and evaluates the methods developed for the module analysis of problems of the closed subject system of answers to questions for the Kazakh language (QA). Analyzing questions in order to obtain the necessary information to determine what is asked and how to respond to them, analyzing questions is one of the most important components of a quality control system. Therefore, we propose new methods for analyzing questions based on two main problems, in particular, on the basis of rules (Rule-based) and on the Hidden Markov model (HMM), based on the integration of the system classification approach, focus extraction and question classifiers, both of which use the relationship of dependence between the words in the question. A comparison of these solutions with basic models is also given. This study also provides a manual summary and annotated data of the gold standard for further research in this area.