基于时间引子的购物网站用户兴趣变化研究  

Study on the change of user interest in shopping website

作　　者：刘一鸥[1] 

机构地区：[1]西安外事学院,陕西西安710077

出　　处：《电子设计工程》2017年第20期112-115,共4页Electronic Design Engineering

基　　金：2016年西安市社会科学规划基金项目(16IN13);2015年陕西省教育厅科学研究项目(15JK2113);2015年陕西省高等教育教学改革研究项目(15J37)

摘　　要：针对购物网站营销的精准化需求,提出一种基于时间效用函数的个人兴趣推荐模型。在该推荐模型中认为人的兴趣是随着时间的变化而不断变化的,由此客户在最近的评分更可以体现出该用户的购物兴趣,并且评分越接近的用户相似度越高。对此,在对相似度进行计算中引入时间效应函数,对不同评分赋予不同的效用值,从而提高相邻相似度的准确性,大大提高购物推荐的准确性。最后通过对比实验验证了上述方案的可行性。Aiming at the precision demand of shopping website marketing, a personal interest recommendation model based on time utility function is proposed. In the recommendation model that interest is changing with the change of time, thus customers in recent score can reflect the user＇s interest in shopping, and score closer and closer to the user similarity is high. In this regard, the time effect function is introduced to calculate the similarity, and the different values are given to different utility values, so as to improve the accuracy of the adjacent similarity, and greatly improve the accuracy of the recommendation. Finally, the feasibility of the above scheme is verified by comparing experiments.

关 键 词：项目聚类 用户兴趣 个性化推荐 时间效用函数 实验对比 

分 类 号：TN0[电子电信—物理电子学]