An Efficient Feature Selection Method Using Hybrid Particle Swarm Optimization with Genetic Algorithm

Abstract

The data mining applications over big data is a challenging task. The main issues of the big data are velocity problem, variety problem and the volume problem. We want to handle large amount of data in the case of big data such as medical data, sensor data, telephonic record data etc. In some cases, the classifier is not good enough and do not work well for data which have many features. Too many features are affects the effectiveness of classifier, some features may be redundant. Too many features goes through the classifier, which will cause increasing the workload of the classifier. In order to solve this problem, we need some optimized feature selection method. In this work proposed an algorithm called Hybrid Particle Swarm Optimization with Genetic Algorithm (HPSOGA). This is a very good feature selection method to find the optimal features for the classification to overcome the draw backs of the classification model. The efficiency of the classification model can be done using this feature selection algorithm through selecting the relevant and the significant features. So it help to obtain improved accuracy within the reasonable processing time of the classifier.