摘要
随着移动终端技术和传感技术的快速发展,如今可以很方便地通过移动终端获得用户所处位置情景信息。目前基于位置推荐的研究已有不少,但对位置情景在推荐系统中的运用方式、重要程度及权重分配上仍有不足之处。针对现有研究的不足,在传统的用户×项目二维推荐模型的基础上,引入位置情景,建立用户×项目×位置三维模型,提出一种针对移动终端环境的混合多维推荐模型,融合了位置情景相似度过滤、协同过滤以及项目相似度过滤三个维度的推荐。实验表明,该推荐模型具有更好的推荐效果。
With the rapid development of mobile terminal technology and sensor technology, it is easy to obtain the location context of mobile users via mobile terminals. Currently there are many researches about location-based recommendation, however, the way, the importance degree and the weight allocation of location context used in the recommended system is not appropriate enough, Aiming at the inadequacy of the existing researches, a three-dimensional model of User ×Item × Position is created by importing the location context based on the traditional two-dimensional recommended model of User × Item, and a hybrid multidimensional recommended model for mobile terminal environment is also proposed. This model integrates location context similarity filtering, collaborative filtering and content-based filtering. The effectiveness of this recommended model is proved in the end.
出处
《计算机应用与软件》
CSCD
2016年第12期202-206,共5页
Computer Applications and Software
基金
国家自然科学基金项目(11272066)
关键词
个性化推荐
LBS
位置情景相似度
协同过滤
项目相似度
Personalized recommendation
LBS
Location context similarity
Collaborative filtering
Item-based similarity
