PERFORMANCE EVALUATION OF VARIOUS DATA MINING CLASSIFICATION TECHNIQUES THAT CORRECTLY CLASSIFY BANKING TRANSACTION AS FRAUDULENT

Abstract

Data mining is a sum of process to find anomalies, patterns, correlations which can assist banks to look for hidden patterns in a group and discover unknown relationship in the data. In this competitive world, governments, private companies, large organizations and all businesses, predict their future plans using various methods of data mining. Banks are an integral part of a country’s economy, contributing to both people and governments. In case of money transaction bank is the most essential media in our country. In this circumstance, some vested peoples gratify their evil task. These individual acts make a big issue for our country’s economy. In this paper, we have discussed about the comparative study on several data mining classification techniques that are generally used to classify suspicious transactions, included Naïve Bayes, MLP (is the particular feed forward method of ANN), sequential minimal optimization (SMO) and decision tree based J48 & random forest algorithm. we have found the random forest algorithm performance is better than others to classify banking transactions.