Mining Top-K High Utility Itemsets without Generating Candidates

In the domain of data mining, utility mining is the new development area. Due to anti-monotonicity property utility mining is the complicated with itemset share framework. There exist several techniques on mining utility with two phase candidate generation approach. However, such approach is inefficient as well as not scalable for huge dataset. In two phase candidate generation approach, scalability issue is detected in case of huge number of candidates. The proposed system is efficiently identifying high utility itemsets without candidate generation. Reverse enumeration tree is introduced to reduce search space by utility upper bound. d2 HUP enables to compute tight bound efficient pruning & directly discovers the HUI in scalable as well as efficient way. Along with the proposed work, system can also discover the top-k itemsets from extracted results. With the experimental results, we have to show that the time required for pattern enumeration is very less using d2HUP algorithm than the existing techniques of candidate set generation.