Novel feature selection method using mutual information and fractal dimension

In this paper, a novel feature selection method using Mutual Information (MI) and Fractal Dimension (FD) to measure the relevance and the redundancy features is presented. The proposed algorithm maximises the relevance and minimises the redundancy of the attributes simultaneously. The new framework allows a more efficient method for the selection of features without using any search technique. The performance of the proposed algorithm is compared with three different feature selection methods on three different datasets. The results obtained confirm the comparable efficiency and effectiveness of the features selected through the proposed algorithm.