Feature selection is a main step in the data classification process, and it is applied to determine optimal features and develop an accurate model. This paper presents a novel top-down correlated data reduction method which is employed by feature selection process. The method removes redundant features by clustering and placing correlated features in a cluster. This operation continues until reaching a certain and desired number of features. The proposed feature selection method resulted in performance rates of 88.21%, 89.41%, 87.64%, and 86.54% in terms of accuracy, sensitivity, specificity, and f-measure, respectively which are highly effective compared to the current state-of-the-art approaches.
Mehdi Ayar (Tue,) studied this question.