摘要
协同过滤技术是推荐系统中运用成功、广泛采用的推荐算法,但其在计算项目相似性需要依靠用户的历史评分数据。基于此,本文提出了利用项目自身属性进行项目分类,然后在类内寻找最近邻居,最后产生推荐。实验结果表明,该算法提高了数据稀疏情况下的推荐质量,同时还提高了推荐系统的效率。
Collaborative filtering is the most successful and extensively used recommendation algorithm in the existing recommendation systems which rely on the users' historical rating data to calculate the project similarity. In view of this, this paper proposes the project categorization by their own properties, and then finding the most adjacent in the class to produce the recommendation. Experimental resuits show that the algorithm not only improves the recommendation quality with insufficient data, but also the efficiency of recommendation system.
出处
《洛阳理工学院学报(自然科学版)》
2015年第3期57-60,70,共5页
Journal of Luoyang Institute of Science and Technology:Natural Science Edition
基金
国家科技支撑计划资助项目(2012BAH20F05)
甘肃省自然科学基金(1310RJZA056)
兰州交通大学青年基金(2013016)
兰州交通大学青年基金资助项目(2011012)
关键词
项目属性
协同过滤
推荐系统
project property
collaborative filtering
recommendation system