Efficient Mining of High Utility Patterns Using Frequent Pattern Growth Algorithm

Data mining aims at extracting only the useful information from very large databases. Association Rule Mining (ARM) is a technique that tries to find the frequent itemsets or closely associated patterns among the existing items from the given database. Traditional methods of frequent itemset mining, assumes that the data is centralized and static which impose excessive communication overhead in case of distributed data, and computational resources are wasted, if the data is dynamic in nature. To overcome this, Utility Pattern Mining Algorithm is proposed, in which itemsets are maintained in a tree based data structure, called as Utility Pattern Tree, which generates the itemset without examining the entire database, and has minimal communication overhead when mining with respect to distributed and dynamic databases. Hence, it enforces the execution to be at a faster rate, means reduced cost and time.