Abstract: Along with great increase in ATM Card transactions, ATM card fraud has become increasingly widespread in recent years. This paper work investigates that the efficacy of applying HMM algorithm is to prevent the ATM card fraud problems. The different techniques & classification methods, i.e. HMM is tested for their applicability in counterfeit or fake detection. The proposed system provides a useful framework to detect & prevent a fraud & to choose the best model to recognize the ATM card fraud risk. This methodology uses the different security layers before entering the pin no. This authentication mechanism is useful while transaction to secure pin no. or cash card by sending OTP to user for verification in case of ATM card transactions. The main objective of this proposed system is that instead of relying on pin no. as a security no. the multiple layers of security is enhanced & implemented by HMM algorithm. This authentication mechanism is useful while transaction to secure cash card from being cloned via skimmed device& providing more security i.e. only by judging person only by its pinno.

Keywords: Hidden Markov Model, Advanced Fraud Detection System.