Mining Top-k Frequent Closed Itemsets in Data Streams Using Sliding Window

Frequent itemset mining has become a popular research area in data mining community since the last few years. There are two main technical hitches while finding frequent itemsets. First, to provide an appropriate minimum support value to start and user need to tune this minimum support value by running the algorithm again and again. Secondly, generated frequent itemsets are mostly numerous and as a result a number of association rules generated are also very large in numbers. Applications dealing with streaming environment need to process the data received at high rate, therefore, finding frequent itemsets in data streams becomes complex. In this paper, we present an algorithm to mine top-k frequent closed itemsets using sliding window approach from streaming data. We developed a single-pass algorithm to find frequent closed itemsets of length between user‟s defined minimum and maximum-length. To improve the performance of algorithm and to avoid rescanning of data, we have transformed data into bitmap based tree data structure.