Keyword Search Based on Lexical Relationships in the Text

The new method of the keywords determining based on finding the relationships between word forms of the English text is suggested in this paper. Existing approaches to the definition of keywords are characterized. The best quality of text processing is achieved by linguistic methods or when their combinations are statistical. A system for automatically determining keywords from natural language text should be developed using the morphological dictionary and syntax rules. In the framework of the general approach to reducing verbal noise in the method it is proposed that it is achieved with the help of formalized operations: the replacement of pronouns with the corresponding nouns; removal of noise connections; removing noise words; withdrawal of stop words. The described operations can be used as additional modules that improve the results of finding keywords for both the developed method for determining keywords of English text and other algorithms for finding keywords.