This paper deals with the description of a complex method for the disambiguation of part-of-speech homonymy in Russian texts. The method is based on the data on syntactic compatibility of Russian words. A method for compiling a similar corpus is discussed.
The author proposes a method for document retrieval with the use of a document pattern. For this purpose, a complex measure for assessing the subject similarity of documents is introduced. A numerical evaluation of the results of a search output for different measures is given.