Analysis Of Inflectional Behaviour In Indian Languages Using Features Extraction Techniques

Abstract

Compared to English, Indian languages are highly inflectional in nature. Features extraction is a challenging task for Indian languages due to multiple reasons. First, alphabets called ‘Akshara’ are coded with Unicode unlike English, which is coded with ASCII. Second, upon inflection, the structure of the root word gets changed to a different format. Words get modified according to their added features as per tense, aspect, and modality. Thirdly, composite characters called ‘yuktAkshara’ are influenced with additive vowels or consonants. With these three prospects, this paper is aimed to address some practical difficulties for stemming root words with experimental reviews. Feature extraction is used in hidden information retrieval, root word stemming, text-to-speech conversion, and semantic analysis of Natural Language Processing. Analyzing features from an inflected word of Unicode language, one can recover the semantic and pragmatic meaning of the written text and further can be used in text-to-speech conversion. This paper discusses various techniques of feature extraction and its applications in some Indian languages.