<![CDATA[Bad credit is one of the credit risk faced by the financial industry and banking. Credit jams can be avoided by means of an accurate credit analysis of borrowers. The accuracy of credit ratings is crucial to the profitability of financial institutions. Improved accuracy of credit ratings can be done by selecting the attributes, because the selection of attributes reduce the dimensionality of the data so that the operation of the data mining algorithms can be run more effectively and more quickly. The purpose of this study is to apply the Particle swarm optimization (PSO) to do the selection of attributes on a Support Vector Machine to improve the accuracy of the accuracy of credit analysis at anugerah.Banyak employee cooperative research has been conducted to determine credit ratings. One of the methods most widely used method of support vector machine. In this study will be used method of support vector machine and will have the attributes using the Particle Swarm Optimization for determining credit ratings. After testing, the results obtained are support vector machine produces accuracy rate 74.74%, 82.03% and AUC values precision value of 0.643. Then the selection of attributes to use, particle swarm optimization in which the attributes which originally numbered 11 predictor variables were selected seven attributes used. The results showed a higher level of accuracy that is equal to 81.36%, 83.08% and AUC values precision value of 0.689. There by achieving an accuracy improvement of 6.62%, and increase the AUC of 0.147. By looking at the values of accuracy and AUC, support vector machine algorithm particle swarm optimization into classification category enough.]]>
Copyrights © 2016
