Feature selection using the hybrid of ant colony optimization and mutual information for the forecaster

Abstract

Feature selection in the forecaster based on artificial neural network is a well-researched problem, which can improve the network performance and speed up the training of the network. In this paper, we proposed an effective feature selection scheme called ACOMI, which utilizes the hybrid of ant colony optimization (ACO) and mutual information (MI). In this method, mutual information between each input and each output of the data set is employed in search process to purposefully guide search direction of every ant in ant system, and the parameter Exploit can adjust the balance between the ability of the cooperation among ants and the inherent ability to exploit. By examining the forecasters at the Australian Bureau of Meteorology, the simulation of three different methods of feature selection shows that ACOMI can reduce the dimensionality of inputs, speed up the training of the network and get better performance. In addition, the performance and cost time can be adjusted by the parameter of Exploit.