Impact of Stemming on Telugu Text Classification

In Text categorization, Information retrieval and document clustering stemming is absolutely necessary especially for morphological rich languages like Indian. The process of stemming is, reducing the inflected or resultant terms to their stem word, root or origin form. However, stemming is a tricky task - particularly for extremely inflected natural languages having a lot of words for the same normalized word form. In Text classification, stemming tries to cut off details like either suffix or prefix of a word and produce basic word. In this paper, we apply various stemming methods on Telugu text classification and ensure the performance of the classifier is effect by stemming. Telugu is suffix oriented language, so we have performed number of experiments on erratically selected Telugu text documents and finally we conceive that the performance of the classifier is improved.