Citation

Kalid, Suraya Nurain and Khor, Kok Chin and Ng, Keng Hoong and Ting, Choo Yee (2014) Effective classification for unbalanced bank direct marketing data with over-sampling. In: Knowledge Management International Conference (KMICe) 2014, 12 - 15 August 2014, Langkawi, Malaysia.

Text Effective Classification for Unbalanced Bank Direct Marketing Data with Oversampling.pdf Restricted to Repository staff only Download (351kB)

Abstract

Direct marketing is an effective tool used by financial institutions such as banks to promote products and services. This study began with analyzing an unbalanced direct marketing data set and subsequently discovered the rare class problem. The problem has prompted unsatisfactory classification rate for an important minority class in the data set. The analysis shows that the problem is caused by the small size of the minority class, overlapping of classes, and noisy data. Further, the incapability of classification algorithms in handling unbalanced data sets has also contributed to the problem. Data-level approaches are normally used to solve or mitigate the problem. The authors evaluated various data-level approaches using the data set and identified SMOTE (an over-sampling technique) as the effective approach to the problem. The authors also managed to identify a set of useful features for classification, instead of using all the features in the data set.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Bank direct marketing, unbalanced data set, rare class problem, SMOTE, data set, big data
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	06 Apr 2015 04:55
Last Modified:	23 Nov 2016 06:49
URII:	http://shdl.mmu.edu.my/id/eprint/6147

Downloads

Edit (login required)