Multi Novel Class Classification of Feature Evolving Data Streams with J48

Abstract

the Data stream classification main issues are infinite length, concept drift, concept development, and feature development. Hypothetically data stream is infinite in length; it is impossible for storing and use all the traditional for training. In the existing system of data stream method researcher tackle on the only two issues i.e. concept drift and concept evolution problem of classification. In the existing system for tackling the issue of feature evolution feature set homogeneous technique was developed and also focus on the novel class detection technique for detecting the novel class at a time, but this method required more time for detecting novel and multi novel class detection. Therefore we used the method for detecting the novel class method for data stream classification, we used J48 classification algorithm for detecting the novel class and reducing the time for detecting the novel class. Finally we compared our result with the existing novel class detection method. KeywordsData Stream Classification; J48 classifier; novel class; features evaluation