A New Intelligent Hybrid Feature Selection Method

摘要

In recent years, the adoption of technologies has grown notably in many domains and resulted in dense amounts of data. processing such large big data became one of the main challenge in the technology society and nictitates intelligent tools to deal with. The feature extraction and selection are main steps in classification systems that aims to lower the complexity, reduce the dimensionality, generalizable models building and most importantly enhancing the performance as well as accuracy rates due to the quality and significance of the selected features. The aforementioned steps often increase the model performance and often enhance the classification accuracy rate. In this paper we introduced a new hybrid feature selection method and evaluate it against ten datasets form UCI repository, experimental results show that the classifier we've adopted to the experiment has achieved better classification accuracy when compared with the other version that used a single feature selection method.