Detecting A Fraudulent Financial Statement And Reducing Internal Fraud, Using Data Mining Techniques

A Fraudulent Financial Statement (FFS) is the Intentional furnishing and/or publishing of false information in it. In this thesis we use Data mining techniques that help to identify fraudulent financial statements and reduce internal fraud, by using Decision trees, neural networks and Bayesian Belief networks. These techniques are expedient, especially when new methods of fraudulent financial statements adapt to the detection techniques."These three techniques are compared using the same data sample with different models, and it shows that BBN outperforms the other two models and achieve outstanding classification accuracy."This thesis also explores a self-adaptive framework (based on a response surface model) with domain knowledge to detect fraudulent financial statements and how to reduce internal fraud. To conclude, this paper suggest that, in an era with evolutionary financial fraud, computer assisted automated fraud detection mechanism will be very effective and efficient with specialized domain knowledge.