Mathematical model of an ontological-semantic analyzer using basic ontological-semantic patterns

Abstract

In this work we propose a mathematical model of a Russian-text semantic analyzer based on semantic rules. We provide a working algorithm of the semantic analyzer and demonstrate some examples of its software implementation in Java language. According to the developed algorithm, the text that proceeds to the input of the semantic analyzer is gradually reduced: some syntaxemes from the analyzed text, in accordance with semantic rules, are added to a queue with priority; then on each iteration the syntaxeme corresponding to the highest priority element is removed from the text. When a syntaxeme is removed from the text, the corresponding element is removed from the queue. Priority definition for a queue element bases on the value of the group priority of the semantic dependence which is described by a semantic rule, as well as on the position in the analyzed syntaxeme corresponding to the queue element.