Hunger Search Algorithm with Optimal Deep Learning Driven Credit Card Fraud Detection and Classification Model

Abstract

With the development of e-commerce websites, persons and financial companies depend on online services for executing its transaction which has performed an exponential improvement from credit card fraud (CCF). The fraudulent credit card transaction gives rise to a loss of massive count of money. The proposal of effectual fraud detection method was essential for reducing the losses suffered by the customer and financial company. The research is complete on several methods for preventing and detecting CCFs. Feature of CCFs role a vital play if machine learning (ML) was utilized to credit card fraud detection (CCFD), and it can be selected accurately. This article focuses on the design of hunger search algorithm with optimal deep learning is driven credit card fraud detection and classification (HSAODL-CCFC) model. The major intention of the HSAODL-CCFC model is to properly distinguish credit card transactions from legitimate or fraud. To accomplish this, the presented HSAODL-CCFC model applies data pre-processing at the initial stage. Next, a new binary version of HSA based feature selection (BHSA-FS) approach was executed for electing feature subsets. Finally, root means square propagation (RMSProp) with deep belief network (DBN) technique was utilized for the identification and classification of the CCFs. The simulation analysis of the HSAODL-CCFC model is tested using 2 benchmark datasets. The comparative analysis reported the promising performance of the HSAODL-CCFC model over recent approaches.