ABSTRACT
The RecSys Challenge 2016 focuses on the prediction of users' interest in clicking a job posting in the career-oriented social networking site Xing. Given users' profile, the content of the job posting, as well as the historical activities of users, we aim in recommending top job postings to users for the coming week. This paper introduces the winning strategy for such a recommendation task. We summarize several key components that result in our leading position in this contest. First, we build a hierarchical pairwise model with ensemble learning as the overall prediction framework. Second, we integrate both content and behavior information in our feature engineering process. In particular, we model the temporal activity pattern using a self-exciting point process, namely Hawkes Process, to generate the most relevant recommendation at the right moment. Finally, we also tackle the challenging cold start issue using a semantic based strategy that is built on the topic modeling with the users profiling information. Our approach achieved the highest leader-board and full scores among all the submissions.
- O. Aalen, O. Borgan, and H. Gjessing. Survival and event history analysis: a process point of view. Springer Science & Business Media, 2008.Google Scholar
- F. Abel, A. Benczúr, D. Kohlsdorf, M. Larson, and R. Pálovics. Recsys challenge 2016: Job recommendations. In Proceedings of the 2016 International ACM Recommender Systems, 2016. Google ScholarDigital Library
- J. Bennett and S. Lanning. The netflix prize. In Proceedings of KDD cup and workshop, volume 2007, page 35, 2007.Google Scholar
- D. M. Blei, A. Y. Ng, and M. I. Jordan. Latent dirichlet allocation. Journal of machine Learning research, 3(Jan):993--1022, 2003. Google ScholarDigital Library
- R. Burke. Hybrid web recommender systems. In The adaptive web, pages 377--408. Springer, 2007. Google ScholarDigital Library
- T. G. Dietterich. Ensemble learning. The handbook of brain theory and neural networks, 2:110--125, 2002.Google Scholar
- N. Du, M. Farajtabar, A. Ahmed, A. J. Smola, and L. Song. Dirichlet-hawkes processes with applications to clustering continuous-time document streams. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 219--228. ACM, 2015. Google ScholarDigital Library
- N. Du, L. Song, H. Woo, and H. Zha. Uncover topic-sensitive information diffusion networks. In Proceedings of the sixteenth international conference on artificial intelligence and statistics, pages 229--237, 2013.Google Scholar
- N. Du, Y. Wang, N. He, J. Sun, and L. Song. Time-sensitive recommendation from recurrent user activities. In Advances in Neural Information Processing Systems, pages 3492--3500, 2015. Google ScholarDigital Library
- J. H. Friedman. Stochastic gradient boosting. Computational Statistics & Data Analysis, 38(4):367--378, 2002. Google ScholarDigital Library
- A. G. Hawkes. Spectra of some self-exciting and mutually exciting point processes. Biometrika, 58(1):83--90, 1971.Google ScholarCross Ref
- C. G. ianqi Chen. Xgboost: A scalable tree boosting system. In Proceedings of the 22th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, 2016. Google ScholarDigital Library
- T.-Y. Liu. Learning to rank for information retrieval. Foundations and Trends in Information Retrieval, 3(3):225--331, 2009. Google ScholarDigital Library
- M. J. Pazzani and D. Billsus. Content-based recommendation systems. In The adaptive web, pages 325--341. Springer, 2007. Google ScholarDigital Library
- F. Ricci, L. Rokach, and B. Shapira. Introduction to recommender systems handbook. Springer, 2011.Google ScholarDigital Library
- P. Romov and E. Sokolov. Recsys challenge 2015: ensemble learning with categorical features. In Proceedings of the 2015 International ACM Recommender Systems Challenge, page 1. ACM, 2015. Google ScholarDigital Library
- J. B. Schafer, D. Frankowski, J. Herlocker, and S. Sen. Collaborative filtering recommender systems. In The adaptive web, pages 291--324. Springer, 2007. Google ScholarDigital Library
- I. Valera and M. Gomez-Rodriguez. Modeling adoption and usage of competing products. In Data Mining (ICDM), 2015 IEEE International Conference on, pages 409--418. IEEE, 2015. Google ScholarDigital Library
- J. Wang, K. R. Varshney, and A. Mojsilovic. Legislative prediction via random walks over a heterogeneous graph. In SDM, pages 1095--1106. SIAM, 2012.Google Scholar
- Z. Zheng, K. Chen, G. Sun, and H. Zha. A regression framework for learning ranking functions using relative relevance judgments. In Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, pages 287--294. ACM, 2007. Google ScholarDigital Library
- Z.-H. Zhou, J. Wu, and W. Tang. Ensembling neural networks: many could be better than all. Artificial intelligence, 137(1):239--263, 2002. Google ScholarDigital Library
Index Terms
- Job recommendation with Hawkes process: an effective solution for RecSys Challenge 2016
Recommendations
A General Rating Recommended Weight-Aware Model for Recommendation System
HCC 2016: Revised Selected Papers of the Second International Conference on Human Centered Computing - Volume 9567In recommendation system, the ratings represent the users' preference and play an important role in recommending items to users. However, the ratings of items may be influenced by many factors, such as time the latest ratings are more able to reflect ...
Cross Domain Framework for Implementing Recommendation Systems Based on Context Based Implicit Negative Feedback
The last decade met a remarkable proliferation of P2P networks, PDMS, semantic web, communitarian websites, electronic stores, etc. resulting in an overload of available information. One of the solutions to this information overload problem is using ...
A Novel Social Event Recommendation Method Based on Social and Collaborative Friendships
SocInfo 2013: Proceedings of the 5th International Conference on Social Informatics - Volume 8238Many social network sites (SNSs) provide social event functions to facilitate user interactions. However, it is difficult for users to find interesting events among the huge number posted on such sites. In this paper, we investigate the problem and ...
Comments