An Efficient Mining Algorithm for Determining Related Item Sets using Classification and Association rules

In the present days, data mining is the advanced research area because it is one of the important step in the knowledge discovery process. This paper presents an experimental study of finding the frequent item sets by classifying the data base transactions into classes by using Decision tree induction based classification and applying Frequent-Pattern (FP) growth on the classes. First, data base transactions are preprocessed by using the pre-processing techniques and those are classified into classes based on information gain. After classifying the transactions into classes, we applied the FP growth algorithm to obtain the frequent or related item sets. This proposed technique is also suitable for heterogeneous data bases. We examined this technique on different types of data bases and by using this technique it have given the accuracy of 96%