介绍传统协同过滤方法,提出一个新的基于情景的多维协同过滤推荐模型。在该模型中,介绍情景的概念;阐述建立基于情景的多维用户模型的方法,并对基于情景的多维协同过滤推荐模型的组成部分进行详细介绍。提出一种计算情景相似度的新算法。基于该新算法,可以得到当前用户的“最近邻”在当前用户所在情景下对项目的评分。 This paper introduces the traditional collaborative filtering method and proposes a new recommendation-based multi-dimensional collaborative filtering model. In this model, the concept of scenario is introduced; the method of establishing multi-dimensional user model based on scenario is explained, and the components of multi-dimensional collaborative filtering recommendation model based on scenario are introduced in detail. A new algorithm for calculating the similarity of scenarios is proposed. Based on the new algorithm, we can get the current user’s “nearest neighbor ” in the current user scenario of the project score.