1Assistant Professor, Silver Oak, College of Engineering and Technology, Ahmedabad, Gujarat
2Assistant Professor, Information Technology Department, Chandubhai S. Patel Institute of Technology, Charusat, Changa, Gujarat, India
*E-mail id: desaiankitb@gmail.com
Feature selection is the important step in overall data mining process. There exist several methods for feature selection. This paper discusses numerous methods of feature selection and their impact on the credit approval dataset. Cost-sensitive feature selection using a decision-tree-based algorithm is proposed (CSattribSelectorC4.5). Sequential minimal optimisation (SMO) is used for classification. The classification results show that the CSattribSelectorC4.5 helps SMO to reduce the overall misclassification cost on the selected dataset.
Cost-sensitive decision trees, Data mining, Feature selection, Weka, CSattribSelectorC4.5
