Performance Examination of Feature Selection methods with Machine learning classifiers on mobile devices

Abstract: Machine Learning is a field which deals with programming computers or mobile device that learns from experience. The field of Machine learning is a common research stream in Computer Science. Machine Learning techniques are helpful in several fields of Computer Science, Information Technology, Mobile computing, e-learning, Bioinformatics, Network Security, Agriculture and Web Document Categorization. Machine learning classification algorithms are used in Image Recognition, Pattern Recognition, Text Classification, Mobile message categorization, Mobile Image tagging applications, Mobile music selection according to user interests, Mobile learning. This work examines the Classification Accuracy of Bayesian classifier and Nearest Neighbor classifier on Mobile device with Android Environment. We present in this work the performance examination of both the classifiers with four feature selection methods of Correlation, Gain Ratio, Information Gain and Symmetrical Uncertainty.

Keywords: PIMS, FIRT, k-NN.