Credit Card Fraud Detection

Abstract

Plastic money is fast replacing currency notes. This has led to cashless travels, which is very convenient. Despite all the merits of such money transactions, there are some demerits also leading to money loss through fraud. Fraud in credit card is increasing manifold. Such frauds are serious issues currently in the industry and cause huge damage. are costly, time-consuming, and labor-intensive tasks. Consumers and financial companies pay very high cost amounting to billions of dollars yearly due to frauds, and fraudsters are on the go to find newer ways and techniques to commit unlawful actions, continuously. Definition of credit card can be given by the physical loss of credit card or loss of sensitive credit card information. As these frauds are increasing very fast we need efficient methods like Machine Learning and Artificial Intelligence techniques to tackle them. Various challenges include unpublished literature and an imbalance classification. Here, in paper we suggest to add a new module of face recognition and detection as a confirmatory test in case of doubtful transactions. Fraud detection are combination of number activities performed to avoid money from being stollen or taken by illegal or untrue instances. In banking, fraud activities involve forging of cheques or misusing robbed or stollen credit cards. In the paper, the focus is on frauds of credit card and to understand how classifier algorithms are performing. We would be considering sampling the data using various algorithms and performing a comparative analysis on the results of classifiers post sampling. Resultant, we hope to retrieve a classification algorithm working best with a particular sampling methodology dependent on the dataset and parameters provided. Lastly, addition of a face recognition model is performed to act as a confirmatory test in case of fraud detection by the classifier.   

Keywords:  Face recognition, class imbalance, classification algorithms, fraud.