ABSTRACT
Data breach is a common problem in the current age of digitalization and paperless transactions. Banking and payment industries are preferred targets for this type of cyber-crime. Both rule based fraud-monitoring systems and predictive modelling systems have failed to provide a fool-proof solution to the problem owing to their respective limitations in its approach and manner of institutional handling of such cases of frauds. This paper proposes a hybrid approach to solve the problems posed by various models and techniques. A comparative analysis of models proposes a loan-blend method addressing the issues. It proposes a methodology of detecting data crimes in Credit Card industry. It compares and contrasts various machine learning algorithms in terms of their accuracy and effectiveness in combating the same. Real time data-sets are used for modelling and testing different algorithms in it and validating the model using confusion matrix. A functional architecture is undertaken for exploration of its uniqueness, relevance and implications.
