The Sequence Association Rules Mining For The E-commerce Personalized Recommendation

With the rapid development of electronic commerce, Shopping on the Internet has become indispensable in people’s daily life. However, the information which the page of electronic business website displays always has some limitation, which can’t display all of the information which the user has interest of at the same time. However,due to a wide variety of goods and the difference of people’s interests, the users have the information overload problem on choosing products. Fortunately, Personalized recommendation technology is one of the most effective means to solve the problem of information overload. By the recommendation technology, the transaction is reached between the user and the electronic sellerquickly and easily.Most of the recommendation technologies now pay attention to the interest of user in one purchase behavior, or they regard all of the users’ purchase records as their eternal interests, which didn’t consider the dynamic change of their interests by time. Sequence pattern discovery is the mature and important technology to solve the problem,which aim to discover the relationship between different purchase affairs. Considering the dynamic change of users by time,this article study to apply the sequential pattern mining to the personalized recommendation better, which include the following :First,through the analysis of rule based personalized recommendation algorithm, this paper found that the recommendation efficiency based on the short rules is higher than that based on the long rules. At the same time, by the analysis ofthe sequential pattern mining, this paper found and prove that the collection of binomial sequential patterns can represent the entire sequence pattern information. So this article applies the binomial sequential patterns to the personalization recommendation.Second,Owing to the low efficiency of existing algorithms in the sequence pattern mining, we developed the simple sequential pattern in this article, which only scans the database once and avoid the scanning of the redundant users. So the algorithm in this paper improves the effectiveness and efficiency ofpersonalized recommendation.Third,however, the sequential pattern mining mainly focus on the interest relationship among different purchase affairs, less of attention of relationship on the same purchase affair. So this paper combine the sequential pattern mining and association rule mining which mainly focus on the interest relationship in the same purchase affairs to the recommendation technology.By the verification of the real Amazon data, the algorithm in this paper improves the effectiveness and efficiency ofthe personal recommendation.