Posted by on 2017年7月3日

[隐藏]

1.Traditional Recommender System

Personalized recommendation system is an important method to solve the asymmetric information problem, such as inefficient information matching caused by the information overload of the supply side. It is very likely that the buyer is unable to gather enough information on the quality of the trading item from the seller when they want to select the items from the similar products. At this Internet and intelligence era, collecting information is becoming more and more cheap. The more complete the user information, the higher accuracy the recommender system can achieve.

The traditional methodologies of recommender system are generally Collaborative Filtering Recommendation Algorithm, including user-based, item-based and implicit semantic models. And preference networks form the basis for collaborative filtering algorithms and traditional recommender systems, which are techniques for predicting new likes or dislikes based on comparison of individuals’ preferences with those of others. (Resnick, P., & Varian, H. R., 1997) Early collaborative filtering algorithm has two core ideas: First, system will find a group of similar users for the specific user, and then select items that can be recommended from the shopping data of these similar users; Second, if the two users like the same item, then the system will recommend one of the original user's favorite items to another user. This latter approach, often referred to as Item-Item collaborative filtering, is an effective recommendation algorithm and a classic algorithm used early in the Amazon shopping system.

Recently, so many methods have been constantly proposed, such as Latent Factor Model (Agarwal, D., & Chen, B. C., 2009), Matrix Factorization (Bell, R. M., & Koren, Y., 2007;Koren, Y., 2010), Tensor Factorization (Rendle, S., & Schmidt-Thieme, L., 2010, Feb.), Collective Matrix Factorization (Singh, A. P., & Gordon, G. J., 2008, Aug.), Factorization Machine (Rendle, S., 2010, December), Multi-armed Bandit (Li, L., Chu, W. et al, 2010, Apr.) and Thompson Sampling (Thompson, W. R., 1933) based on Bayesian Methods (Agarwal, D. et al, 2013). Ensemble of Implicit Semantic Model and Probabilistic Graphical Model, especially Topic Model (Wang, C., & Blei, D. M., 2011), Recommender System and Information Retrieval, especially Learning To Rank techniques (Chapelle, O., Le, Q., & Smola, A. , 2007) have also been successful so far. A typical system application of Ensemble Learning is Facebook's modeling of the "News Feed": learning high-level features through the Gradient Boosted Decision Tree (GBDT), and finally with a simple Linear models learn a combination of high-level features. And there are also method using control theory (Jambor, T., Wang, J., & Lathia, N. (2012, April)) and Collaborative deep learning (Wang, H., 2015, Aug.; Wu, Y., 2016, Feb.) to extract the high level features and to improve the accuracy of score prediction. With the success of AlphaGo, deep reinforcement learning has become a very hot research direction. However, the algorithm of this direction is not far away from the real application in recommender system. The main obstacle is that there is a gap between product and technology to be solved.

2.Personal Recommender System

Any recommendation system that targets the user's preferences might meet the problem of low exploration (March, J. G., 1991; Yi, X. et al, 2014, Oct.). And existing recommendation techniques tend to recommend similar items all the time and especially not efficient for low-loyalty customers. Therefore, personalized recommender systems, which take consideration of the personalized information of customers, will be the trend of research on recommender system. Xu, X., et al (2014) present a scheme taking consideration of the non-functional aspects of products (ease of use, UI design, power consumption etc.). Cho, Y. H., Kim, J. K., & Kim, S. H. (2002) suggested a personalized recommendation methodology based on a variety of data mining techniques such as web usage mining, decision tree induction, association rule mining and the product taxonomy. Bao, J., Zheng, Y., & Mokbel, M. F. (2012) present a location-based and preference-aware recommender system that offers a particular user a set of venues within a geospatial range with the consideration of both user preferences and Social opinions. GK, K. K. B., & Sheela, N. (2016) combine the social network (three social factors, personal interest, interpersonal interest similarity and interpersonal influence ) with E-commerce site to develop an efficient recommender system. From all of these we can take use of modern techniques of big data and machine learning especially deep learning to collect more information of specific agents of the contract of business and through the analysis such as data interconnection, user portraits and so on to make more wise recommendations and decisions.

As social networks such as Facebook, Weibo have been so popular and the sharing economy mode business such as Uber can give the clue of location information of customers. This project is planned to taking use of the social network and location information and the techniques such text mining and sentiment analysis to improve the performance of personalized recommender system.

3.Reference:

Cho, Y. H., Kim, J. K., & Kim, S. H. (2002). A personalized recommender system based on web usage mining and decision tree induction. Expert systems with Applications, 23(3), 329-342.
Chicago
Bao, J., Zheng, Y., & Mokbel, M. F. (2012, November). Location-based and preference-aware recommendation using sparse geo-social networking data. In Proceedings of the 20th international conference on advances in geographic information systems (pp. 199-208). ACM.
Resnick, P., & Varian, H. R. (1997). Recommender systems. Communications of the ACM, 40(3), 56-58.
Liu, J. G., Chen, M., Chen, J. C., Deng, F., Zhang, H., Zhang, Z., & Zhou, T. (2009). Recent advances in personal recommender systems. International journal of information and systems sciences, 5(2), 230-247.
Aboody, D., & Lev, B. (2000). Information asymmetry, R&D, and insider gains. The journal of Finance, 55(6), 2747-2766.
Marwala, T., & Hurwitz, E. (2015). Artificial Intelligence and Asymmetric Information Theory. arXiv preprint arXiv:1510.02867.
Malkiel, B. G., & Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The journal of Finance, 25(2), 383-417.
Agarwal, D., & Chen, B. C. (2009, June). Regression-based latent factor models. In Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 19-28). ACM.
Koren, Y. (2010). Factor in the neighbors: Scalable and accurate collaborative filtering. ACM Transactions on Knowledge Discovery from Data (TKDD), 4(1), 1.
Bell, R. M., & Koren, Y. (2007). Lessons from the Netflix prize challenge. ACM SIGKDD Explorations Newsletter, 9(2), 75-79.
Wang, C., & Blei, D. M. (2011, August). Collaborative topic modeling for recommending scientific articles. In Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 448-456). ACM.
Chapelle, O., Le, Q., & Smola, A. (2007). Large margin optimization of ranking measures. In NIPS workshop: Machine learning for Web search.
Rendle, S. (2010, December). Factorization machines. In 2010 IEEE International Conference on Data Mining (pp. 995-1000). IEEE.
Chicago
Singh, A. P., & Gordon, G. J. (2008, August). Relational learning via collective matrix factorization. In Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 650-658). ACM.
Rendle, S., & Schmidt-Thieme, L. (2010, February). Pairwise interaction tensor factorization for personalized tag recommendation. In Proceedings of the third ACM international conference on Web search and data mining (pp. 81-90). ACM.
Li, L., Chu, W., Langford, J., & Schapire, R. E. (2010, April). A contextual-bandit approach to personalized news article recommendation. In Proceedings of the 19th international conference on World wide web (pp. 661-670). ACM.
Agarwal, D., Chen, B. C., Elango, P., & Ramakrishnan, R. (2013). Content recommendation on web portals. Communications of the ACM, 56(6), 92-101.
Thompson, W. R. (1933). On the likelihood that one unknown probability exceeds another in view of the evidence of two samples. Biometrika, 25(3/4), 285-294.
Jambor, T., Wang, J., & Lathia, N. (2012, April). Using control theory for stable and efficient recommender systems. In Proceedings of the 21st international conference on World Wide Web (pp. 11-20). ACM.
Wang, H., Wang, N., & Yeung, D. Y. (2015, August). Collaborative deep learning for recommender systems. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1235-1244). ACM.
Wu, Y., DuBois, C., Zheng, A. X., & Ester, M. (2016, February). Collaborative denoising auto-encoders for top-n recommender systems. In Proceedings of the Ninth ACM International Conference on Web Search and Data Mining (pp. 153-162). ACM.
Mikolov, T., Karafiát, M., Burget, L., Cernocký, J., & Khudanpur, S. (2010, September). Recurrent neural network based language model. In Interspeech (Vol. 2, p. 3).
March, J. G. (1991). Exploration and exploitation in organizational learning. Organization science, 2(1), 71-87.
Xu, X., Dutta, K., & Datta, A. (2014). Functionality-Based Mobile App Recommendation by Identifying Aspects from User Reviews.
Yi, X., Hong, L., Zhong, E., Liu, N. N., & Rajan, S. (2014, October). Beyond clicks:

4.Others

用机器学习的知识写一个博客推荐系统,是不是需要用户来对看过的博文评分?

Posted in: 未分类
6 views
1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading...

Comments

  1. 读张小龙后的一点想法 – 饱蠹阁baoduge - […] 之前由于申请,看来一些关于Recommender System,很感兴趣。 […]
  2. 一些学习资料的整理 – 饱蠹阁baoduge - […] RECOMMENDER SYSTEM SURVEY […]

Leave a Reply

返回顶部