A New Discretization and Pattern Selection Method For Classification in Data Mining Using Feedforward Neural Networks

Abstract

This paper proposes a new supervised mean wise discretization algorithm and pattern selection method. A new supervised mean discretization algorithm automates the discretization process based on the mean value of discretizing attribute in each target class. The results obtained using this discretization algorithm show that the discretization scheme generated by the algorithm almost has minimum number of intervals and requires smaller discretization time. A new pattern selection method proposed in this paper is to select the discretized patterns with various features based on pattern disparity for training the feedforward neural network which leads to the improvement in convergence speed and classification accuracy. The efficiency of the proposed algorithm is shown in terms of better discretization scheme and better accuracy of classification by implementing it on six different real data sets. Keywords: supervised discretization; classification; data mining; pattern selection; backpropagation training algorithm; multilayer feedforward neural network.