Root based stemmer for telugu script

Abstract

In this paper, a new stemmer has been proposed named as “Root based stemmer”. This stemmer is strictly based on Dravidian script. Stemming can be used to pick up the effectiveness of information retrieval. In proposed Root based stemming technique, each and every token is compared against with all the words of a valid root words dictionary until a match is found. Then extract the matched string or substring from a token and identified as valid root. The present work is aimed to build dictionary based stemmer to extract valid root words for Indian languages especially for Telugu and compare the results with existing stemmers.