METHOD OF SEARCHING TERM INTERPRETATIONS FOR DOMAIN DICTIONARIES, USED FOR DEVELOPING SOFTWARE

In this paper, a method of forming definitions of terms for a vocabulary of a subject domain using existing explanatory dictionaries is proposed. It is shown that with a combined search for terms and their interpretations, it is possible to find about ten percent of definitions, which is clearly not enough. A method of automated search for the interpretation of terms is proposed, involving the use of existing explanatory dictionaries. A mathematical model of the subject domain dictionary entry is proposed. A mathematical model of an explanatory dictionary entry is proposed, taking into account the headword, a variety of interpretations of the word, litters and stable phrases. A mechanism has been developed for extracting definitions of a term from an explanatory dictionary depending on the structure of its dictionary entry. An algorithm for automated search for definitions for single-word terms has been developed. An algorithm has been developed for the automated search for definitions for verbose terms, based on the selection of nouns from the term. A mechanism for assessing the quality of possible interpretations, depending on the occurrence of terms from the subject domain, is proposed. A mechanism has been developed for the choice of definitions, when the terms from the vocabulary of the subject domain and the explanatory dictionary are coinciding incompletely, which is based on the procedure of term decomposition, the search for partial interpretations and the synthesis of the resulting interpretation. The software developed that allows to organize the search for interpretations of terms both in local explanatory dictionaries (previously loaded into the system), and in online dictionaries. The expert's task includes the evaluation the interpretations found and possibly editorial correction of them. Experimental evaluation of the effectiveness of the use of a software product showed a reduction in the expert's working time compared to the “manual mode” by approximately four times.