Recent weighted maximal frequent itemsets mining

Abstract

Frequent Itemsets Mining is a major task in data mining that is used for discovering the itemsets that occur frequently in a large database. But in real world, every item cannot be given same significance. And also items that are frequent very recently are very important ones to be mined. Frequent Itemsets Mining is incorporated with these two factors to give a new technique called Recent Weighted Frequent Itemsets Mining. Recent Weighted Frequent Itemsets Mining is an extension of Frequent Itemsets Mining task that is used for mining the frequent itemsets that satisfy the weight and recency constraints. This paper focuses on mining maximal recent weighted frequent Itemsets. Maximal Recent Weighted Frequent Itemsets is a compact representation of Recent Weighted Frequent Itemsets. This representation is very much useful when the database size is very large and memory is an issue. Number of maximal recent weighted frequent itemsets will be very much lesser than the number of recent weighted frequent itemsets for a given dataset. The time required for mining and memory required for storing is very less for maximal frequent itemsets when compared with frequent itemsets. The results of the proposed method show an improvement in the time complexity and also the memory space required for mining.