An Enhanced Hui Miner Algorithm to Retrieve Optimum Number of High Utility Itemsets

Utility Mining is a rising research domain in the data mining and it discovers high utility itemsets in the data sets. The utility can be any user defined choice and to assign this utility value is based on the attributes of the dataset. For example in a transaction database it can be a price of purchased item, profit of item or quantity of items purchased. An itemset is known as high utility itemsets if its utility value is greater than minimum utility threshold. Utility mining is used in many applications like inventory control, retail store or super market management and marketing. Most of the existing algorithms generate candidate set to calculate high utility itemsets. Candidate item generation is more difficult and tedious task. In order to avoid this, high utility itemsets can be retrieved without candidate generation. An algorithm named as HUI Miner algorithm retrieves high utility items without candidate generation method. In this work, HUI Miner algorithm is enhanced with a utility function and set of effective pruning strategies. In this work utility threshold is calculated automatically by using utility function. Performance of enhanced HUI Miner algorithm is compared with the existing HUI miner algorithm. Different sizes of datasets are used for experimentation. The average size of the transaction is 10. Results proved that the enhanced algorithm reduces the upper bounds of high utility itemset. Hence it saves time and memory.