International Journal of Computer Applications
Most of the data mining algorithm focuses on frequent patterns, few algorithm emphases on rare items, but rare items also have importance, for example, network intrusion detection, where among various normal connections the authors need to detect the rare malicious connections. Classification of such a non-uniform data set is a challenging issue. Most classifiers perform poorly in such a data set. Realizing the importance of rare class classification, in this paper, they propose a classification algorithm (CBMR Algorithm) that is based on association rules mined by MS Apriori approach and is capable of classifying rare classes.