Efficient Predictions on Asymmetrical Financial Data Using Ensemble Random Forests

Abstract

The technological advances in the areas of Big Data and machine learning have led to many useful applications in the financial industry. However, the success of these technologies depends on the analysis of useful information. The financial data is often asymmetrical in nature. It is the nature of information that is crucial in making financial decisions. It is often used to detect the financial frauds, predict the market trends, marketing financial products, and various other use cases. In this work, we are proposing that the ensemble random forests will be able to make better predictions on the asymmetrical financial data. We are taking two cases for making the predictions—one, predicting the customers who will buy the term deposit and two, credit card fraud detection. In both cases, the ensemble random forests were compared with the logistic regression and demonstrated with the results where the random forests performed better than the logistic regression.