Designing Of Personalization Recommendation System In Mobile Commerce

With the rapid development of mobile communication, especially with the operation of 3G technology (third generation mobile communication technology) in China, accessing to network by mobile is becoming increasingly popular. The high pace of modern society makes people expect to complete the process of business activities while moving through a wireless network, mobile commerce has thus emerged. Mobile commerce can more easily provide unique personalized service for each user, which is the biggest advantage of mobile commerce. Meanwhile, the conflict between the information overload problem and the high cost of information flow also makes it necessary for businesses to implement better personalized service.With perspective of consumer behavior, the paper discusses the contact between consumer buying decision process and to personal service, and provides effective personalized service solutions by designing a personalization recommendation system.First, the paper describes recommendation as a process of understanding users'interests, reducing optional object spaces, and helping users to make decisions. It studies how personalization recommendation system effect consumers'decision making process and results of decision-making. It also discusses factors that affect the utility of recommendation system from three perspectives. And proposals are given based on deficiencies in current application of personalization recommendation system. Then, from the perspective of consumer purchase decision, incorporating weighted combination methods, combining respective advantages of two recommendation techniques with wider application which are content-based filtering and collaborative filtering, the paper mixes these two recommendation techniques together to form a new kind of weighted combination recommendation system. The new system combines user preferences and commodity features together, achieves high performance and accuracy, and alleviates the cold-start problems and sparse problem in recommendation systems to some extent. Finally, through two parts of data experiments using MovieLens data sets provided by GroupLens Research, the paper inspects and verifies the recommendation results in different amounts of information, and considers the influence on algorithm performance by different nearest neighbor set sizes. Test results show that the recommendation system is sensitive to user preferences, can handle large changes of project properties, and has high recommendation accuracy.Aim to resolving contradictions between large number of product information and personalization of consumer demand, the paper improves weighted combination recommendation algorithm, proposes a personalization recommendation method of product information, and has some reference value on the domestic implementation of personalized service in mobile commerce.