中文摘要：

传统的推荐算法一定程度上降低了网络消费者的搜索成本,但难以实时提供消费者满意的推荐服务,也忽略了用户偏好动态转移性。为了提高电子商务系统的推荐质量,从用户偏好的行为特征入手,建立了网络用户的兴趣特征提取模型,并设计了相应的推荐算法。通过对用户兴趣特征提取模型的检验和用户兴趣度矩阵的建立,依据与目标用户偏好相似的邻居用户对商品的兴趣程度预测用户对未浏览商品的兴趣度,并选择兴趣度值较高的N个商品推荐给用户。实验结果表明,在用户偏好动态转移的情况下,所设计的推荐算法的推荐精度和推荐效率明显提高,提高了网络用户的满意度。

英文摘要：

To some extent,the traditional recommendation algorithm reduced consumer＇s online search cost,but it couldn＇t provide satisfactory recommended service for Web consumers timely,and the dynamic metastatic of user preferences had been ignored.In order to improve the recommendation quality of e-commerce systems,based on the behavioral characteristics of the user preferences,this paper established a model on feature extraction of user interests according to the dynamic characteristics of network consumer preferences,and designed a corresponding recommendation algorithm.Users＇ interest model of feature extraction was true and built user interests matrix.By the interestingness of neighbors whose preferences were similar to the target users to predict the users＇ interestingness on those who had no navigation experience in some products,and then recommended the top N products with higher interestingness to Web users.The experimental results show that this method can improve the recommendation efficiency and improve accuracy obviously,and the online consumer＇s satisfaction.