Call For Paper Volume:4 Issue:6 Jun'2017

A New Method Based on Clustering and Feature Selection for Credit Scoring of Banking Customers

Publication Date : 25/02/2016



Author(s) :

Seyedeh Maryam Anaei , Mohsen Moradi.


Volume/Issue :
Volume 3
,
Issue 2
(02 - 2016)



Abstract :

One of the issues which must be taken into account by credit policy makers in banking industry is risk management. Credit risk is one of the most serious risks faced by banks. This risk may bring about negative consequences such as customers’ inability or unwillingness to fulfill their obligations to the bank. To manage and control credit risk, the use customer credit classification system is inevitable. Such system puts customers in appropriate classes based on available data and records .Credit scoring generally aims to provide an accurate prediction of customer competency. To this end, multiple statistical techniques plus artificial intelligence have been used. However, most models are not able to provide multi-class categorizations using two-class data. Nevertheless, a customer is assessed based on some degrees of goodness or badness. Therefore, after removing data noises through clustering, it is usually attempted to use a multi-class support vector machine (SVM) to classify new data. Feature selection and parameter adjustment is performed via genetic algorithm. Compared with other scoring models, the proposed model has been able to improve classification accuracy for German and Australian credit datasets


No. of Downloads :

8


Indexing

Web Design MymensinghPremium WordPress ThemesWeb Development

A New Method Based on Clustering and Feature Selection for Credit Scoring of Banking Customers

February 11, 2016