Recommendation is one of primary ways to addressing the problem of Internet information overload, and has been applied to many areas including E-business. Although there exit different kinds of recommendation algorithms and systems, we are still faced some key challenges which have not been well solved for constructing more intelligent and more robust recommender systems,and the involved research is still the hot point of intelligent information processing. We describes the background and significance of the studies of recommendation techniques,and then introduces main recommendation algorithms including collaborative filtering recommendation algorithms,content based recommendation algorithms,graph structure based recommendation algorithms,and hybrid recommendation algorithms,moreover,discusses the advantages and disadvantages of existing algorithms,and finally concludes main evaluation methods and main issues, improvements and future research directions.%推荐是解决互联网信息过载的主要途径之一,已被广泛应用于电子商务等多个领域.尽管已存在多种推荐算法,建造出更加智能、更加鲁棒的推荐系统仍面临诸多尚未解决的难题,推荐方法的研究仍是智能信息处理的研究热点.文章首先阐述了推荐方法的研究背景、研究意义,之后分别介绍了协同过滤推荐算法、基于内容的推荐算法、基于图结构的推荐算法和混合推荐算法,分析了各类算法的优点与不足,最后总结了主要的评价方法以及面临的主要问题,提出了改进的方法和未来可能的研究方向.
展开▼
