Dependency algorithm have been broadly used to build a community recommendation system from the time when they distinguishes and shares the collective knowledge and occurrence. However it easily falls into the ambush of Mathew consequence, which have a tendency to recommend popular product and that's why the popular product item become increasingly less popular. Underneath this circumstance, the majority of the product items in the recommendation list are already common to customers and because of the performance would seriously decreases in finding exact item. To overcome this issue the prediction mold is built for shopping dataset. The online shopping dataset have been considered for cleaning using regex and mean weighted average vector is proposed. For selection of attributes the algorthim called PCA and Recursive feature elimination has been used and compared their accuracy. Dependency is proposed that can be recommended items to customer using a community build. Using the community mining techniques has been applied like degreeness, closeness, betweeness property of nodes is used to find the loyal customer in the community.

No abstract

No abstract

No abstract