A Unified Approach to Fraudulent Detection

With the increase in demands and price of goods and services, fraudulency has caught a great height. But, it can’t be prohibited completely in the first stage. The detection of fraud has attracted continuous attention from academia, industry and regulatory agencies, and it is a challenging task for the researchers to develop a fraud detection framework. Starting from the late 1900s, ‘Benford’s law’ has served this purpose well. Abruptly, within a decade of its application lots and lots of fraudulency started getting seized. Later on, this law was used for detecting fairness of the elections, forensics, finances, etc. This article proposes a formula specifically derived from Zipf’s law that can detect fairness and fallacies in datasets involving forensics, finances, elections, and similar socio-economic issues. Unlike Benford’s law, our proposed formula is not dependent on any sort of observations, rather it is backboned by rigorous proof. Finally, we have done a comparison analysis between Benford’s law and our proposed formula graphically. All the data sets used by us have been rigorously studied, and many fitting tests have been applied to them.