An Advanced Multi Class Instance Selection based Support Vector Machine for Text Classification

Abstract Machine learning techniques is most commonly used technique in text mining. Support Vector Machine (SVM) is a most useful supervised learning technique for text classification. In this paper we proposed advanced Multi Class Instance Selection based support vector machine (AMCISSVM) to increasing efficiency of support vector machine. The proposed algorithm is compared with Multi Class Instance Selection (MCIS) and Neighborhood Property based Pattern Selection (NPPS) algorithm. The advanced MCIS has shown high accuracy to multi datasets. These experimental datasets are retrieved from UCI machine learning repositories.