Abstract:
We have presented an approach that generates frequent itemsets from data stream. The itemsets are compressed and then stored in the memory. The decision to whether or not compress an itemset is based on the utility of the itemset. In this chapter, the utility of an itemset is defined in terms of the amount of memory saved by its compression. Beside this, we have presented an approach to dynamically generate the value of minimum support threshold based on the data in the data streams. It avoids having a fixed minimum support threshold throughout the data stream. Since the value is generated from the latest elements in the data stream, it suits to be an appropriate measure to separate the frequent itemsets from the non-frequent ones.
