Money transfer on internet is happening very fast. Money applications for payment of salary have been made on the internet. Due to which cases of fraud are also increasing. İn this paper, we have made a little effort to remove this fraud from the hidden markov model. Here, we have used K-means algorithm. The Kmeans algorithm has been used for symbolic and frequency transactions. Starting is normal behavior with HMM account. Any transaction is considered fraudulent if the transaction does not go through the HMM process. Here is an example of real world bank, in which the experiment has been validated by simulation and the model used. İf we look at the old data, then this model works well at low medium frequency.

Keywords : Fraud, Fraud Detection, K-Means, Hidden Markov Model.