Text Categorisation using Semantic and Discriminative Selected Feature

Computer Engineering

Classification of data is very useful technique in data mining, but classification in massive amount of data, data takes much time for classification, which is a big issue. So reduce the features used for classification is an essential task for making time required by system. Previously organize set of classes and documents, automatically categories data based on feature matching. In proposed system find out patterns which increases speed as well as accuracy of text categorization in this digital era. For that selection best features is very important. It will improve performance text categorization and reduce the interpretation. For good organization here first use kullback-Leibler (KL) Divergence and Jefferys Divergence as binomial representation that matter type I and type II errors of Bayesian classifier. Then introduce a new a divergence measure called Jeffreys Multi Hypothesis (JMH) Divergence for multiclass classification. Develop two efficient method of feature selection first is maximum discrimination and second is maximum discrimination x2. Further to improvement the bias capacity develop feature selection algorithms by weighting each distinct feature. The promising results of extensive experiments demonstrate the effectiveness of the proposed approach.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.