摘要
近年来O2O电子商务模式兴起并迅速发展。针对移动互联网领域的信息过载和个性化服务推荐问题,通过将移动用户的位置信息引入到协同过滤的推荐过程,提出一种基于位置信息改进的协同过滤推荐算法。该算法首先通过计算用户与项目的距离对项目集进行预过滤,以项目被评分用户的交集作为计算项目相似度的基础,对预过滤项目集进行偏好预测。实验表明,该算法能有效大幅减少推荐过程中的计算量以改善推荐的实时性,并能对新项目和新用户作出推荐。
O2O e-commerce model rises and rapidly develops in recent years. Directed at he information overload and personalized service recommendation problem in the mobile Internet field, this paper introduces the user's location information into the collaborative filtering recommendation process, proposes an improved location-based collaborative filtering algorithm. Firstly, by calculating the distance between the user and the item to pre-filter the item set, the intersection of two items rated hy users as the hasis for computing similarity, to conduct preference prediction for the set of pre-filtered item. Experimental results show that the algorithm can significantly reduce the amount of calculating in the recommendation process to improve real-time, and can make recommendations on new projects and new users.
出处
《系统工程》
CSSCI
CSCD
北大核心
2015年第12期121-125,共5页
Systems Engineering
基金
国家自然科学基金资助项目(71371077)
国家自然科学基金重大资助项目(71090403
71090400)
关键词
O2O模式
位置信息
协同过滤
推荐系统
冷启动
O2O Mode
Location Information
Collaborative Filtering
Recommendation System
Cold Start