Evaluation of Credit Card Threats using Incremental Learning Approach

Abstract

Credit card is the well accepted manner of payment in financial field. With the increasing number of users across the globe, risks on usage of credit card has also been raised, where there is danger of stealing of credit card details and committing frauds. Incremental methods are desirable in recent machine learning applications such as financial problems like credit card threat assessment since amount of data and information is intensifying over the time. Scale up in learning can be achieved by updating classifier as and when training data becomes available. A smart technique known as ensemble technique has become popular, in which multiple classifiers are united in such a way that correct decisions are amplified and incorrect ones are discarded. Major focus of ensemble based techniques is diversity of classifiers that leads to reduction in misclassification. This paper presents ensemble based technique named as NIK algorithm, which handles credit data efficiently and finally distinguishes the bad customers from faithful ones in more accurate way. General Terms Pattern Recognition, Financial Security, Machine learning