Implementing frequent item set mining by overcoming over-scan problems

Abstract

This paper over comes the problem of over-scan problems which are occurring frequently in the frequent itemset mining. In case of improved apriori we improve the efficiency and decrease the time lapse but it is unable to solve the over-scan problem efficiently. We now propose a systematic approach for the immediate solving of the over-scan problems by implementing the RP Tree on spark framework. The adaptation of this approach will be very useful in the formation of the tree of frequent patterns and also for the visualization of the frequent-1-itemsets. This is mainly to overcome the over-scan problems in the previous improved algorithm.