An Effective Approach for Newspaper Article Classification using Multi-Class Support Vector Machine in Comparison with Binary Classifier to improve Accuracy

Abstract

Aim: An efficient approach to classifying newspaper articles using a multi-class Support Vector Machine. Materials and Methods: Accuracy stands as result for the classification of text analysis. Factual texts merely attempt to inform, whereas virtual texts try to amuse or combative readers by inventive language and imagination. The rate of correct classification of novel texts is low and classification occurs in the areas of text analysis and classification of multiple articles. The binary classification and the separation of data points into classes. The multiclass SVM is used for splitting the multiple into severely binary classification. The Novel Text Classification is checked by sample size (N = 42) Support Vector Machine obtained with G-Power taking value equal to 80%. Results: Accuracy is the outcome, Support Vector Machine accuracy rate is 82.71%, which is relatively higher than the Binary Classifier (BC) with 71.48%. Significance value accuracy becomes 0.101 (p>0.05). Conclusion: SVM works and gets more accurate than the Binary Classifier. And this research is evaluated to predict accuracy for a system that is proposed Support Vector Machine is higher than existing comparison utilizing Binary Classifiers.