Paper Details: Downloads: 495
Serial Number: P1111008002
Title: Implementing Automated Prediction Systems for Credit Scoring
Authors: Alaa M. Elsayad
Abstract: The aim of credit scoring system is to model or predict the probability that consumers with certain characteristics are to be considered as potential risks. Until recently, the decision to grant credit was based on human judgment to assess the risk of failure to pay. From the literature it has been found that data mining models such as support vector machine (SVM), multilayer perceptron neural network (MLPNN) and decision tree (DT) can help decision makers to improve in this domain. SVM, MLPNN and DT with their remarkable ability to derive meaning from complicated or imprecise data, can be used to extract patterns and detect trends that are too complex to be noticed by either humans or other conventional techniques. This paper evaluates the performance of SVM with polynomial kernel (POLY-SVM), MLPNN with pruning parameters and DT with Chi-squared automatic interaction detection (CHAID) algorithm in the prediction of consumer’s credibility. The dataset contains information related to the sociodemographic and financial characteristics of consumers from real world credit bank. Known sets of risky and creditworthy consumer data were used to train the models to categorize new cases. The purpose was to determine an optimum classification model with high predicting accuracy for this problem. Experimental results demonstrate the effectiveness of CHAID-DT model in predicting credit scoring with higher accuracies than other techniques.
Keywords: Credit scoring, data mining, support vector machine, neural network, decision tree.
Journal/Conference: International Journal of Automatic Control and System Engineering
Volume: 10
Issue: 1
Submission Date: 2/18/2010 12:00:00 AM
Review Date: 3/24/2010 12:00:00 AM
Publishing Date: 5/19/2010 12:00:00 AM
Article Downloads: 495
Download:

Facebook