Ontological-semantic text analysis and the question answering system using data from ontology

Today, with an avalanching increase in information, the task of developing the systems that allow user to quickly search desired information in large text volumes is becoming more and more urgent. An example of such system is the question answering one. In the work we describe an architecture of such system which work is based on utilizing data from an ontology. We propose an algorithm for automatic update of the ontology basing on use of an expert system and ontological rules for logical inference. We also describe an ontology with the structure based on object-oriented model and describe the functions that are used to update the ontology and extract data from it. We describe the way to update the ontology and to modify the stored data using the rules stored in the ontology. For writing ontological-semantic rules we use the Drools expert system that utilizes the PHREK algorithm for fast pattern matching. We analyze the issues of using Apache Spark system for distributed implementation of the algorithm.