The Application Of Time Weighted Collaborative Filtering Algorithm In Electronic Commerce

With the development of Internet technology and data processing technology, people have entered a new era——from lack of information era to the era of information overload. This change also brings a huge problem to people:on the one hand, for those users, how to find their own needed interested information has become a challenge? on the other hand, for the information producer,how to sell their goods and attract users eyeball, but also a problem which need to consider. Based on the above issues, recommendation systems emerged and became a communication bridge between the above problems. The purpose of recommendation system is to provide more convenient and more efficient service for users and business information, it not only can help the users to find something interesting, but also can make businesses sell their goods, show to the users who need them. The role and significance of recommendation systems is very important in electronic commerce website. Therefore, research on application of recommendation algorithm in e-commerce, will has great research significance.Collaborative filtering algorithm based on the project does not consider the time weight, thus proposed adding time consider factors to improve item collaborative recommendation algorithm, designed the time weighted collaborative filtering e-commerce recommendation system model, the book website as the application background, designed a shopping website recommendation model and the that was verified.Firstly,study and compare the user-based and item-based collaborative filtering algorithm,it is proposed that the time factor is added to the recommendation system and the time weight is calculated, which can be closer to the user’s latest interest state.The similarity calculating formula based on project and score was improved, and the formula of time weighted similarity was used instead of the original formula.Moreover, to reduce the time cost, the algorithm combines the clustering method of data mining, and the project space is firstly clustered to reduce the computation time and space expenses, and the computation efficiency is increased.Finally, the improved "time type based on the project content and scale of the weighted collaborative filtering algorithm" to carry on the simulation experiment. Experimental results show that the improved algorithm in the mean absolute deviationand average time consumption are better than the original item based collaborative filtering algorithm. It is proved that the improved algorithm is efficient.Finally,we design a recommendation system model based on electronic commerce. With the online bookshop website as the actual application background of the recommended model, we design a personalized recommendation model of shopping website, the main content includes: the main function of system structure design, model analysis, the recommended process analysis, each function module design and database design etc.