Text Categorization of Marathi Documents using Modified LINGO

In the 21st century we have seen tremendous growth in the users who are accessing internet. At the same time users across the universe are giving preferences to the regional languages while accessing internet. It is seen that there is a huge scope for doing categorization in regional languages, considering the fact that search engine has worked on many regional languages. Moreover, super-fast acceleration in the information technology has driven the variety of applications of text categorization. We have chosen Marathi as a regional language for automatic text categorization. Categorization defines various classification and clustering techniques and Label Induction Grouping (LINGO) is one of these algorithms used for categorization of Marathi text documents. This uses Single Value Decomposition (SVD) for dimension reduction. We are proposing a modified LINGO algorithm which is considering morphology of Marathi text and Principle Component analysis (PCA) for improving results.