Abstract

Credit card frauds have become significant owing to the rise of latest technologies and the global superhighway of communication. Credit card fraud identification models are necessary for any bank or financial institution to reduce the loss. Several methods are developed for the identification of credit card frauds. To solve this problem, the proposed system designed a Time-Varying Inertia Weight based Dragonfly Algorithm (TVIWDA) with Weighted Feature based Support Vector Machines (WFSVM) for classifying the normal/fraud behavior features. In this proposed work, initially, credit card dataset is taken as an input and feature selection is performed based on Time-Varying Inertia Weight based Dragonfly Algorithm (TVIWDA). According to the selected features, Weighted Feature based Support Vector Machines (WFSVM) approach is utilized for credit card fraud detection. A brief set of experiments were performed to highlight the betterment of the proposed model and the obtained simulation values ensured the better outcome of the proposed model over the compared methods.