1) taking POIs as the item, doing item recommendation
2) taking images posted by users as the content of items using CNN technology
3) taking PMF as the basic model and taking the probabilistic model as constraints (让从CNN中学出的image vector 与用户vector以及item vector产生关系，如：用户自己的vector同自己post的image的vector尽可能近；poi的vector同与之相关的item尽可能相近)
4）CNN ref： http://cs231n.github.io/convolutional-networks/ and http://www.jeyzhang.com/cnn-learning-notes-2.html
| 文献题目 | 去谷歌学术搜索 | ||||||||||
| What Your Images Reveal: Exploiting Visual Contents for Point-of-Interest Recommendation | |||||||||||
| 文献作者 | Suhang Wang | ||||||||||
| 文献发表年限 | 2017 | ||||||||||
| 文献关键字 | |||||||||||
| POI as item, image as content of item, PMF, CNN, | |||||||||||
| 摘要描述 | |||||||||||
| The rapid growth of Location-based Social Networks (LB- SNs) provides a vast amount of check-in data, which facili- tates the study of point-of-interest (POI) recommendation. The majority of the existing POI recommendation methods focus on four aspects, i.e., temporal patterns, geographi- cal influence, social correlations and textual content indica- tions. For example, user’s visits to locations have temporal patterns and users are likely to visit POIs near them. In real-world LBSNs such as Instagram, users can upload pho- tos associating with locations. Photos not only reflect users’ interests but also provide informative descriptions about lo- cations. For example, a user who posts many architecture photos is more likely to visit famous landmarks; while a user posts lots of images about food has more incentive to visit restaurants. Thus, images have potentials to improve the performance of POI recommendation. However, little work exists for POI recommendation by exploiting images. In this paper, we study the problem of enhancing POI recom- mendation with visual contents. In particular, we propose a new framework Visual Content Enhanced POI recommenda- tion (VPOI), which incorporates visual contents for POI rec- ommendations. Experimental results on real-world datasets demonstrate the effectiveness of the proposed framework. | |||||||||||
