Research And Implementation On Timely Recommendation System Of E-Commerce Web Site Based On Apriori

People have been turning their traditional way of shopping to the way of online shopping gradually because of the high effect and convenient network. As the e-commerce recommendation system not only can retain customers effectively but increase the sales and competitiveness of e-commerce companies as well, it has become an important role in the IT technology research of e-commerce and has drawn many researchers'attentions. At present, although there are so many researches on e-commerce recommendation system, it is still difficult to give a timely recommendation system model limit to large trading volume and data mining algorithms.In this paper, a timely recommendation system of e-commerce web site is designed and implemented using the implicit and efficient data acquisition together with the data mining technology of association rules.The AprioriTidStr method is proposed in this paper as an improved algorithm of Apriori. The proposed method can realize frequency itemsets search with the operation of the Tid which is the identifier of each affair correspond to its data item sets in the database. This method avoids the scanning the database repeatedly so that it can reduce the mining complexity of association rules. And on this condition, we design a complex mining method which combines both the real-time online mining and the regular manual mining. This can make the recommendation system update recommended goods real-time when recommending. Multi-functional mixed recommendation is realized in the proposed system such as recommendation of customer acts history, recommendation of current view and buying, and recommendation of the popular goods on the web site.Finally, we give a brief performance analysis of the proposed system. Also we put forward where the system needs to improve and the next step in the further study.