Performance and Evaluation of Different Kernels in Support Vector Machine for Text Mining

Abstract

Text mining is the subfield of data mining. Text analysis or mining is enormously growing field for research simultaneously to artificial intelligence and data mining. Unstructured data or multimedia data is being constantly generated via social media Web sites, call center logs, blogs, and so on. Therefore, the explosion of textual data in a social media network is overwhelming. Text analysis or mining is widely being used to determine meaningful and noteworthy information for the huge amount of unstructured multimedia or heterogeneous data. In this paper, we compare the performance of text classification through support vector machine using multimedia data. This text analysis uses the technique of different fields like predict, machine learning, information, visualization, and natural language processing. The use of support vector machine for learning text nonlinear classifier is determined here. The reason for using SVM is that its performance is more accurate as compared to other soft computing tools and algorithms. In this paper, different kernel tricks have applied with nonlinear classifier for classification of text data mining. The results of proposed experiments predict that support vector machine with radial basis function achieves the highest overall accuracy.