The selection feature for batik motif classification with information gain value

Abstract

Features in the classification process have an important role. Classification of batik motif has been done by using various features such as texture features, shape features, and color features. Features can be utilized separately or combined between features. The problem in this research is how to get the potential feature to classified the motif of batik. The feature combination causes enhancement in the number of features causing dataset size changes in the classification process. In this research will be done the selection of features to the combination feature of texture and feature of shape from batik motifs. The feature selection process uses the information gain value approach. The feature selection is done by calculating the value of the information gain of each feature of texture and feature of shape. The value of the information gain will be sorted from the highest information gain value. Ten features with the highest information value will be the selected feature to be processed in the process of batik image motif classification. The classification method using in this research is an artificial neural network. The neural network consists of three layers, that is input layer, hidden layer, and the output layer. The data from selection feature processed in the artificial neural network. The result of this study shows that the accuracy of the process of batik motif classification with a combination feature of texture and feature of shape is 75%. The addition of feature selection process to batik motif classification process gives an increase of 12.5% to the yield an accuracy of 87.5%.