中文摘要：

多维信息推荐在推荐的过程中考虑情景因素对用户行为的影响，动态捕捉用户兴趣在不同情景下的变化，向用户提供更加个性化、智能化的推荐结果。文章分析多维信息推荐的相关内容，提出了输入情景化这一新的概念，构建了基于降维的输入情景化多维信息推荐系统模型，研究了基于降维的输入情景化多维信息推荐算法，并通过实验研究的方法验证了新算法的高效性与优越性。

英文摘要：

Multi-Dimensional Information Recommendation （MDIR） takes the influences of the contextual factors on user behaviors into account in the recommendation process, dynamically catches the changes of the user interest in different contexts, and provides users with more personalized and intelligent recommendation results. This paper analyzes the related content of MDIR, pro- poses the new concept of contextual pre-fihering, and based on the dimension reduction-based contextual pre-filtering, constructs the model of the multi-dimensional information recommendation system and develops the multi-dimensional information recommenda- tion algorithm. Finally, the paper uses the method of experimental study to verify the efficiency and advantages of the new algo- rithm.