Classifier Ensemble for Imbalanced Data Stream Classification

Abstract

The data streams in various real life applications are characterized by concept drift. Such data streams may also be characterized by skewed or imbalance class distributions for example Financial fraud detection, Network intrusion detection etc. In such cases skewed class distribution of the stream increases the problems associated with classifying stream instances. Learning from such skewed data streams results in a classifier which is biased towards the majority class. Thus the classifier built on such skewed data streams tends to misclassify the minority class examples. In case of some applications like financial fraud detection the identification of fraudulent transaction is the main focus because here misclassification of such minority class instances will result in financial loss. Similarly in case of many other real life data stream applications the misclassification costs associated with minority class instances are higher and they need proactive treatment. In this paper we present our preliminary work where in we propose a method which makes use of k nearest neighbours and oversampling technique to balance the class distributions. Experimental results show that the approach shows good classification performance on synthetic and real world data sets.