A Combined Model Of Decision Trees Algorithm For Predicting Users’ Shopping Behaviors

With the popularity and rapid development of Internet, an explosive amount of data is brought out in the fields like banking, financial services, e-business. The mag-nitudes of users and brands grow profoundly by progression in electronic commerce. According to the records of user’s past shopping and properties of merchandise, per-sonalized information recommendation system should be become one of hot-spots in E-commerce.the feature also has an important value in the search of the internet page, is emerging data mining research field. Personalized recommending may be valued for users as well as sellers.That the extraction of hidden predictive information from large data-set, is a pow-erful new technology with great potential to help companies focus on the most impor-tant information in their data warehouses. In this paper, we introduce the basic theory of e-commerce recommendation service and several classic algorithms, and then pro-filed the data-set provided by e-commerce corporation alibaba and constructed a series of features from it. Due to the nonlinearity and mutuality of the features, we intro-duce the Random Forest(RF) and Gradient Boosting Decision Tree(GBDT) based on Decision Tree to figure out this problem, which both using a large of decision trees to approach the target value, and meantime reduce the system errors. In addition, the fea-tures are selected though the importance calculated by RF. Experimental result shows good precision and recall value, and compare these two algorithm by performance and properties.