Abstract. This paper describes the software package FEST, which includes a universal program for morphological analysis of scientific and technical texts, MORPH, and several other programs generating data for MORPH. This data includes the morphological tables of a specific input language belonging to the permissible class of inflectional and agglutinative languages and a dictionary. The programs included in the FEST package generate the input language data for the MORPH program using formal descriptions of morphology and vocabularies created by a human expert who knows the language. The analysis strategy is based on an alternation of left-to-right and right-to-left analysis order. The dictionary of the input language contains stems rather than lexemes or word-forms, and consists of several dictionaries, each containing stems of the same length. The stems in the dictionary are accompanied by the grammar information, allowing all the word-forms of the input text to be recognized. The analysis strategy, the structure of the morphological tables and vocabularies enable morphological analysis of all word-forms with stems from dictionary.

Key words: the software package for morphological analysis, formal descriptions of morphology, formal descriptions of lexemes, morphological tables generation, vocabularies generation, results of morphological analysis (description and example).

ACM Classification Keywords: I.2.7. Natural Language Processing – Text analysis.

Link:

УНИВЕРСАЛЬНАЯ СИСТЕМА ПРОГРАММ МОРФОЛОГИЧЕСКОГО АНАЛИЗА НАУЧНО-ТЕХНИЧЕСКИХ ТЕКСТОВ НА ФЛЕКТИВНЫХ И АГГЛЮТИНАТИВНЫХ ЯЗЫКАХ

Надежда Мищенко

http://www.foibg.com/ijitk/ijitk-vol06/ijitk06-1-p05.pdf