A Comparative Assessment of Data Mining Algorithms to Predict Fraudulent Firms

The process of data mining is helpful in discovering meaningful patterns in historical or unstructured data in order to make better business decisions. It helps in creating a better marketing strategy and also helps in risk management, fraud detection, etc. In this study, we put forward a comparative analysis of data mining models for fraud detection. The goal of the analysis is to find the best model which gives high accuracy and is less compute-intensive. We have implemented Decision Trees, Linear Support Vector Machines, RBF Kernel Support Vector Machines, K-Nearest Neighbor, Artificial Neural Network and logistic regression classification models. Further, we have implemented PCA and Ensemble techniques to improve the accuracy of the model and decrease the computational complexity of the models.