Credit Risk Prediction using Clustered Classification

Research areas:

• Volume 3,Issue 5, March 2015

Year:

2015

Type of Publication:

Article

Keywords:

Credit Risk, Decision Tree, Logistic Regression, Support Vector Machine

Authors:

Sedigheh Ghanbari; Saeid Pashazadeh; Hossein Bevrani

Journal:

IJAIM

Volume:

3

Number:

5

Pages:

247-253

Month:

March

ISSN:

2320-5121

Abstract:

Since the granting credit facilities by banks faced many problems such as customer credit risk, it essential for banks that use advancedvalidation customerssystems. The purpose of this study is classification of banks customers who received credit facilities. In this research the process of data mining accomplished on customer credit data that is available on the information system of bank. At first, required data were collected and preprocessing operation applied on them. Then classification techniques including logistic regression, C4.5 decision tree and support vector machine with three different approaches: 1) classification without selecting features and clustering, 2) classification by applying selecting features algorithm and 3) classification by clustering data were used over preprocessed data and the results were compared. Results show that classification based on clustering approach provides higher accuracy in predicting credit risk in comparison with two other mentioned approaches

Full text: IJAIM_416_Final.pdf