期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2011
卷号：8
期号：3
出版社：IJCSI Press
摘要：Lexical resources are very crucial for any NLP research as well as applications. The richness of a Lexical resource depends not only on the coverage of words but also on the coverage of their morphological forms. As keying in all word forms(morphology) for each word can be monotonous and mechanical exercise, hence automatic generation of the word morphology should be important agenda for Lexicon creation tools, in general and more specific for Indian languages, which are morphologically very rich as compared to English. In our ongoing effort of developing Hindi to Marathi Machine Translation (MT) system, we have developed a rule based framework for generating morphology for a given root (stem) word. The framework relies on rules for generating morphology. We have devised a generic format of rule, which caters for necessary grammatical information as required for generating word morphology for specified language. The framework can generate both inflectional as well as derivational morphology of a given word s.t. satisfaction of rule constraints. The rule management, rule selection and choice of selecting correct word forms are some of the important features of the framework.
Loading...
