摘要
社会化标签能够直接反映用户兴趣和商品特征,因而可用于个性化推荐系统中。使用标签进行推荐时,需要将其排序,而现有标签云中的标签排序都是按标签被标记次数或字典顺序进行降序排列的,这些排列方式未考虑用户个性化的需求和标签的时效性。提出了一种个性化标签云中的标签排序算法,通过用户自身的兴趣与挖掘出的用户潜在兴趣进行标签排序,从而构建个性化的标签云,并根据用户兴趣的变化定期对标签云进行更新,可以更好地发现电子商务网站中用户感兴趣的商品。实验对比结果表明,个性化标签云中的标签排序算法能够有效的提高商品推荐质量。
Since a social tag reflects users＇ interest and product characteristics directly, it can be used for personalized recommendation system. Before recommendation the social tag allows sorting either in descending order of taggings or in lexigraphical order. These sorting methods takes into no account users＇ personalized demand and tags＇s time-sensitivity. This paper puts forward a tag sorting algorithm in personalized tag cloud, based on the users＇ present and predicted interests to sort tags and construct a personalized tag cloud. The tag cloud will be updated on a real-time basis according to users＇ interests changing, thus facili- tating discovery of goods users are interested in on e-commerce websites. The experiment result shows that tag sorting algorithm in personalized tag cloud can improve the recommendation quality effectively
出处
《沈阳航空航天大学学报》
2011年第1期46-50,共5页
Journal of Shenyang Aerospace University
