Research Of Item Selecting Algorithm Based On Hypothesis Test And Confidence

With the rapid development of electronic management system,huge amount of operational data is generated every day.Adjusting data mining algorithm in business is becoming more and more voluable.A fundamental data mining problem in business and other application is ranking items with respect to some notion of profit based on historical transactions.The difficulty is that the profit of one item not only comes from its own sales,but also from its influence on the sales of other items,i.e.,the “cross-selling effect”.Researchers has proposed several algorithms to solve this problem.Profset was one of them,however,this model suffered from some deficiencies and pointed as ineffective.The main work of this paper is as follows:In this paper,based on hypothesis test,new measure to discover the patterns,called “valid itemset” was proposed,based on confidence,new step called itemset degeneration was added,according to these imporvement,a new algorithm called “x-profset” was proposed.Sales transaction data from a supermarket and synthetic data are used to demonstrate the effectiveness of the model against a heuristic for product selection based on product-specific profitability and the profset model.x-profset keeps a 2%-6% leading compared with PROFSET at several different item selectiong numbers.Experiment on supermarket data show the measure of “valid itemset” is effective.Besides,based on history transaction data,x-profset is able to provide item profit weight for its own profit and corss-selling profit,which is helpful for related business decision making,experiment is conducted to show that.