Selection of online Features and its Application

Abstract

Selection of Online Feature is significant important concept in data mining. Batch learning is the mostly used learning algorithm in feature selection. Instead of Batch learning, online learning is most efficient and scalable machine learning method. Most existing system studies of online learning should access the data related to features. But accessing all data becomes a problem when we deal with high dimensional data. To avoid this limitation we proposed system in this online learner allowed to operate a classifier having fixed and small number of features related data. But the significant challenge Selection of online features (SOF) is how to construct accurate prediction for a data using a small number of operative features. To develop novel Selection of Online Feature algorithms to perform a various tasks of Selection of Online Feature by using semi supervised and supervised with unlabeled and label data for full input and partial input. Hence it provides integrity and scalability to the data storage system efficiently and users will be accessing the data through online.