标题:
基于兴趣导航的农业科技信息协同过滤推荐方法A Collaborative Filtering Recommendation Algorithm Basing on Preference Navigation for Agricultural Science and Technology Information
作者:
王文超, 宋金玲, 黄立明, 刘爱勇
关键字:
用户兴趣, 协同过滤, 单用户类型, 多用户类型, 相似度User Preference, Collaborative Filtering, Single User Type, Multiple User Type, Similarity
期刊名称:
《Computer Science and Application》, Vol.5 No.6, 2015-06-26
摘要:
由于推荐技术在农业科技信息服务领域的应用处于起步阶段，而传统推荐技术的针对性又存在不足，针对这些问题，提出了基于兴趣导航的协同过滤农业科技信息推荐技术。对于用户类型比较单一的网站，首先根据信息的浏览量选择目标用户的候选兴趣信息，然后根据其他用户对候选信息的评分确定最终的推荐信息。对于用户类型多样化的网站，首先根据目标用户的历史浏览信息及评分确定用户兴趣导航信息，然后根据导航信息确定导航用户集，再通过导航用户集的项目评分确定相似用户集，最后根据相似用户对其他信息进行评分估计并进行推荐。
Although the recommendation technology has been used in agricultural science and technology information service website, it is still obvious that the recommendation result lacks of persona-lization. In order to solve this problem, a collaborative filtering recommendation method basing on preference navigation for agricultural science and technology information is proposed. For the website with single user type, the candidate preference information are selected for the target user firstly according to the browsing quantity, then the recommendation information is determined according to the ratings of the other users. For the website with various types of users, user preference navigation information is chosen according to the browsing history and rating of the target user firstly, the navigation users are determined basing on user preference navigation information secondly, then similar users are determined through the ratings of navigation users, finally predict ratings of other items and recommend according to the similar users.