Investigation on Frequent Itemset Mining using Enhanced Utility and Graph Mining Techniques

Abstract

The Life styles of people have have been changed due to the globalized and digitized marketing on internet world. The internet and e-commerce has changed the way of marketing, selling products and services. Nowadays, the competition among online retailers have become more intensive. Therefore more and more business are trying to gain competitive advantages by using ecommerce to interact with customers. Today industries have moved their focuses from products and sales to customer oriented marketing. The customers behaviour pattern has become important issue of marketing because of heavy competition in the market place. Analysis of customer taste and making of substantial profit have become the twice objective in retail trade. The Shopping of online products is increased due to the revolution of online purchase. Success in business have triggered the need to have updated data and smart mining. This research work is dedicated to this task by selecting, researching, reapplying and enhancing a few existing mathematical models, algorithms and frameworks which are explained below: Utility mining method has definitely proven to be more effective than the frequent mining method in online retail business. The numerical correlation technique blindly takes into account only the total number of items sold. But the technique of utility mining does not undermine this value and considers only the number of item sold (frequency) but examines the utility value of each itemset. One of the major objectives of the frequent mining technique is to discover the preference of customer for an itemset. This research proposes work that involves the combination of the frequency value and the product values of the items in the itemset determine the profit value. The profit value can be optimized by applying the simplex method. The most efficient itemset can be discovered by incorporating the correlation co-efficient technique. This technique is easily applicable and economically feasible at every stage of determining the utility of each items