Analysis of Term Weighting schemes in Vector Space model for text classification

Abstract

The term weighting system (TWS) is an important component for the text matching system whenever the vector space model is employed for information retrieval from text sources. This work reports an innovative way of term-weighting approach to enhance the performance of classification. With respect to text categorization approaches, the term weighting system that we present in this study has the highest accuracy. We examined several weighting - schemes, weight information - gain, SVM, and the current technique's performance parameters with K-NN and Naïve- Bayes technique. A variety of term-weighting techniques (TWM) in conjunction with Information-Gain, SVM, K-NN, and Naïve-Bayes techniques have been used for analysis on Amazon data collections.

URN:NBN:sciencein.jist.2023.v11.469