A Hs-hybrid Genetic Improved Fuzzy Weighted Association Rule Mining Using Enhanced Hits Algorithm

ABSTARCT Earlier the uninteresting rules can be shortened through the fuzzy weighted association rule mining with enhanced HITS algorithm that satisfies downward closure property as a consequence of assigning weights to items manually, which can reduce the execution time. In this FWARM there are two main issues of weight calculation, the foremost one is that the algorithm may not find out customers to describe the suitable membership function all the time, as a result it leads to spend lot of time and increases the profit loss. The next reason is that the customer’s wants to purchase items based on their requirements which are not to be fixed. Thus some more mechanisms are required to adapt the membership functions to these changes automatically and the demand decides on the number of optimum fine-tuning constraints. Also they should find out all predetermined membership functions of all items. To achieve this goal the genetic algorithms are often utilized to obtain a set of appropriate membership functions according to the designed fitness function. The main drawback of GA approach is making the new vector from the existing parent vectors. Also GA is not well suited for fine-tuning constraints and it is essential to incorporate local search methods into GAs. To overcome this problem in proposed work, genetic local search with the support of harmony search with fuzzy weighted association rule mining with enhanced HITS algorithm were performed. The main advantage in HS is from all existing vectors it makes a new vector. This will increase the flexibility in finding the better solutions. The experimental results shows that the proposed genetic based fuzzy weighted association rule mining with enhanced HITS algorithm is more effective when compared with the existing FWARM with enhanced HITS algorithm.