Application And Research On Data Mining In E-Commerce Recommendation System

With the popularity of the Internet, e-commerce has become an increasingly important sales model, e-commerce system provides consumer with more and more goods but the consumers needn't have to go to the shop. As more and more goods can be chosen and the e-commerce site structure has become more complex, the consumers are often lost in the information space of a large number of goods and can not be successful in identifying their needs of commodities. In the competitive environment, a good product recommendation system can provide the users with the goods they needed in order to effectively retain customers and increase enterprises'sales force and competitiveness.Product recommendation system has good prospects of development and application in the e-commerce field; it has become an important research technique in e-commerce application. But with the development of e-commerce, product recommendation system also faced a series of challenges, such as the efficiency and precision and so on. We analyze and research the techniques used in e-commerce recommendation system to resolve the problem it faced.First of all, we analyze the characteristics and importance of data mining and web mining. Secondly, we introduce and analyze e-commerce recommendation system, and then explain their work flow. We also give a simple model of e-commerce recommendation system, and illustrate the various stages of e-commerce recommendation system work flow and key technologies from data preprocessing, pattern discovery, pattern analysis and pattern application. Finally, we change the practical application of the Apriori algorithm following its character so that it can be efficiently applied to improve the recommendation system, and gives full play to its role in e-commerce recommendation system.There are still many shortcomings in the developed Apriori algorithm in the article and it should be further improved, particularly in the accuracy and efficiency. How to improve the efficiency of the association rules algorithm is currently the important issue in the study of various algorithms. With the in-depth analysis and research to the algorithm, we believe that user-oriented e-commerce personalized recommendation service will be more efficient.