Enhancement of classification techniques using principal component analysis and class imbalance handling methods in credit card defaulter detection
by Abhishek Agarwal; Amit Rana; Neeta Verma; Karan Gupta
International Journal of Forensic Engineering (IJFE), Vol. 5, No. 1, 2021

Abstract: The following research reveals the significance of modified classification in estimating new trends. Rigorous evaluation of different classification algorithms viz. logistic regression, decision tree, K-nearest neighbour (KNN) and Naive Bayesian are explored. These findings forecast the finest techniques for discovery of potential defaulters. Our motive is to compare the performance measures between original dataset and original dataset on which principal component is applied. Different algorithms can be compared on the basis of various criterions such as accuracy, precision, F1-score, recall, ROC. We proceeded by applying a general data imbalance handling technique such as smote technique and near miss technique. A comparison is then drawn between the modified dataset with the principle component analysis applied and the imbalance in the original dataset being corrected with the help of under sampling and oversampling. The comparison helps us identifying the best among dimensionality reduction and data imbalance handling techniques on the chosen dataset.

Online publication date: Fri, 03-Sep-2021

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Forensic Engineering (IJFE):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com