International Journal of Computer Applications
Classification can be defined as a target function which maps attribute value of objects to predefined class. One objective is to divide the objects into proper group and other objective is to predict the class of unknown records. Bayesian classifier classifies and predicts the class of objects on the basis of posterior probability based on some prior probability. Earlier paper does not handle the effect of correlated attributes on the performance and the accuracy of classifier. In this paper, a novel approach using association rules is defined to predict the class of unknown records even if the attributes are correlated.