Morphology plays a crucial role in the working of various NLP applications. Whenever we run a spell checker, provide a query term to a web search engine, explore translation or transliteration tools, use online dictionaries or thesauri, or try using text-to-speech or speech recognition applications, morphology works at the back of these applications. We present here a novel computational tool HinMA, or the Hindi Morphological Analyzer, based on the framework of Distributed Morphology (DM). We discuss the implementation of linguistically motivated analysis and later, we evaluate the accuracy of this tool. We find, that this rule based system exhibits extremely high accuracy and has a good overall coverage. The design of the tool is language independent and by changing few configuration files, one can use this framework for developing such a tool for other languages as well. The analysis of Hindi inflectional morphology based on the Distributed morphology framework, its implementation in the development of this tool and integration with NLP resources like Hindi Wordnet or Sense Marker Tool and possible development of a word generator are interesting aspects of this work.
展开▼
