Application And Research Of Decision Tree Technology Based On Rough Sets In Electronic Commerce

With the development of internet technology, electronic commerce is changingthe traditional management pattern in the social economy. In particular, it completelychanges the relations between the enterprise and the customer. In order to obtainrevenue by the smallest investment, electronic commercial enterprises need to findtheir most valuable customers by doing the customer value analysis. According to theanalysis results, they can carry out special promotions and personalized services tomaintain these customers.The main research direction of this article is to apply data mining technologybased on the rough sets to the customer management system for electronic commerce,giving the structure model by using decision tree method based on the rough sets,researching customer value in electronic commerce, withdrawing the useful classifiedrules from the mass data, and then carrying on the customer management for theenterprise of electronic commerce to provide the intelligent decision-making.The main research results include:(1) The value measurement model for customer is constructed by unifying thecustomer current value (customer purchase situation) and the customer latent value(customer interest and etc.).(2) The studies of data mining model based on the rough sets theory is conductedby combining the rough sets theory and the decision tree technology in the data miningtheory. It proposes an improvement of ID3 algorithm--decision tree algorithm basedon attribute dependence theory of the rough sets, and further enhances efficiency of thealgorithm.(3) The practical application is realized by applying the theory to the state-levelproject "Business Intelligent Analyzing Platform for Retail Trade", which is owned byZhong Ke Ying Tai Company in Qingdao. In this project, the article carried on real project analysis for the customer value of electronic commerce by using decision treetechnology based on the rough sets, and got effective rules for classifying customervalue of electronic commerce.