Evaluation of pertinence of linguistic descriptors in information retrieval systems
Larysa Chala, Юлия Юрьевна Харитонова · Eastern-European Journal of Enterprise Technologies · 2015
The possibility of using acronyms as linguistic descriptors for the classification of the analyzed electronic texts is considered in the paper. The proposed approach is implemented using a two-step procedure.In the first stage, acronyms are extracted from several text documents of the field under consideration, followed by preparation of specialized acronyms dictionaries. Search results are sorted to remove impertinent pairs "acronyms/definitions", and then the alignment of the letters, contained in acronyms, with the definition words is carried out.In the second stage, the modified metric DeMT, which allows to determine the pertinence definition of the acronym in the document under analysis is applied. Herewith, the adapted dictionary, created in the first stage of the process under consideration is used. The modified metric , which takes into account contexts can be successfully adapted to the problem of evaluating the pertinence of linguistic descriptors.