AN OPTIMIZED METHOD FOR FINDING HIGH UTILITY ITEMSETS USING FUZZY TAIL NODE TREE (FTNT)

Utility mining is finding the most important profitable items in a large transaction database. It identifies the importance of items using different values. In existing, two steps for finding high utility itemsets: first generates large number of candidate itemsets; then identifies high utility itemsets from generated candidates by doing additional scan in original transaction database. In view of the fact that large number of candidate itemsets generated and requires more memory space. As a final, it decre ases the mining efficiency also it doesn't work with minimum utility threshold value. To overcome this problem, we proposed a new method fuzzy based tail node tree structure (Fuzzy Tail - Node Tree) to find the utility itemsets. The fuzzified quantity inform ation can able to get from transaction database in order to reflect the fuzzy degree of purchased quantity. The performance of FTNT was evaluated in comparison with different types of datasets. The investigational result shows that proposed system is more effective than existing methods in terms of execution time and memory space under minimum utility threshold value. It achieves two orders of scale faster than the state - of - the - art algorithms on dense dataset, and more than one order of magnitude on sparse datasets