Abstract:
In this work we consider a problem of text semantic indexing from positions of text classification problem with lexical ontology units. We offer to use this approach for objects identification problem and to improve quality of information retrieval. Basic differences of classification problem with using lexical ontology units from the text indexing task and from WSD task are formulated. A new method of classification is offered named as OntoKlass. Formal statement of the problem is made. Engineering realization of the method is executed.
Keywords:artificial intelligence, computer linguistics, text classification, lexical ontology, text mining.