摘要
采用用户-景点-线路三部图来描述用户的行为,通过改进的随机行走算法给用户推荐合适的旅游线路,可以提供准确的旅游线路推荐并有效地解决新的线路难以推荐的问题。通过对景点的聚类,减小了数据稀疏性对推荐带来的影响并避免了过拟合问题。实验结果表明,与传统的方法相比,本文提出的算法具有较好的排序准确度,特别是对稀疏度较高的用户,优势更明显。
In order to recommend the travel package to users,a user-place-itinerary tripartite graph was introduced to describe the behavior of the user,then refined random walk algorithm was elaborated to predict the preference of users.The algorithm could give good recommendation and can also effectively solve the new-item recommendation problem.It also reduced the impact of sparsity and avoids overfitting by clustering of places through interest.Compared with traditional methods,experimental results demonstrate that the proposed algorithm has good sorting accuracy and other more obvious advantages especially for the sparser users.
出处
《网络新媒体技术》
2012年第3期42-48,共7页
Network New Media Technology
基金
863重大项目课题-融合网络业务体系的开发(2011AA01A102)
科技支撑项目课题-支持增强型搜索功能的三屏融合服务运行平台(2011BAH11B04)
中国科学院战略性先导科技专项子课题-未来网络架构研究与边缘设备研制(XDA06010302)
关键词
推荐系统个性化旅游随机行走三部图
Recommender System
Personalized
Tourism
Random Walk
Tripartite Graph
