The Research On Association Rules Of The Products

People’s material life and cultural life are enhanced by the coming of the newcentury, esp. the information technology is improved too. Enterprise can abtainmore fruitfull, detailed, multidimensioned datebase followly. The exist of thisdatebase laids a solid foundation for making the strategic planning, marketingprogram and human planning be more scientized, retionalised and targeted. DateMinning was born at the historic moment, which is based on datebase, throughMathematical model, and searches the rules, inadditionally, is a scientific method forstuding consumer preferences and behavior.According to the function, date mining is devised as characteristics distinguish,association rules, classification forecast, cluster analysis, gaffer analysis and so on.Among them, association rules is the focus in computer and mathematics. Apriori isthe classical algorithm, the procedure is concise and explicit, however, the efficiencyis low. In daily life and production, the datebase is multidimensioned, morevaluable significance association rules are important, which can hence the Overallsales profit. This paper puts the emphasis on the association rules, because everyproducts possesses brand attributes and category attributes, at the same time, everycustomer could buy many of one products, so, the paper take comprehensiveweighted association rules in consideration, which contends products number andprofit and is based on layered association rules. In this paper,the main contentsinclude:⑴Introduce the concept and procedures of Date Mining, summary the meritand demerit of classic algorithm.⑵The category association rules ia estamateed by this paper through applyinglayered association rules on based on datebase, to improve the efficiency. On top ofthat analyse the brand association rules through taking the number and profit of the products as the comprehensive weighted association rules, which based on the gettedcategory association rules.⑶A ccording to the improving method, this research makes a testing experiment,using a retail enterprise transaction database. According to the profit, interest andbuy ratio, the results demonstrate the superiority about the comprehensive weightedassociation rules based on layered association rules.