Benchmarking regression algorithms for income prediction modeling. (October 2016)

• Record Type: Journal Article
• Authors: Kibekbaev, Azamat; Duman, Ekrem
• Is Part Of: Information systems. Volume 61(2016)
• Page Start: 40
• Page End: 52
• Publication Date: 2016-10
• Subjects: Regulation -- Income prediction -- Regression techniques; Database management -- Periodicals; Electronic data processing -- Periodicals; Bases de données -- Gestion -- Périodiques; Informatique -- Périodiques; Database management; Electronic data processing; Periodicals; 005.7
• DOI: 10.1016/j.is.2016.05.001
• Languages: English
• ISSNs: 0306-4379

Abstract

Abstract: This paper aims to predict incomes of customers for banks. In this large-scale income prediction benchmarking paper, we study the performance of various state-of-the-art regression algorithms (e.g. ordinary least squares regression, beta regression, robust regression, ridge regression, MARS, ANN, LS-SVM and CART, as well as two-stage models which combine multiple techniques) applied to five real-life datasets. A total of 16 techniques are compared using 10 different performance measures such as R2, hit rate and preciseness etc. It is found that the traditional linear regression results perform comparable to more sophisticated non-linear and two-stage models. Highlights: This paper focuses on modeling income for the purposes of setting credit card limits. It reviews the performance of six linear and five non-linear methods as well as five approaches combining OLS and non-linear models. Those 16 techniques were applied to five real-life datasets. The resulting models were compared using 10 different performance measures.