ABSTRACT
This paper provides an overview of the approach we adopted as team Lunatic Goats for the ACM RecSys Challenge 2017 [7]. The competition, organized by XING.com, focuses on a cold start job recommendation scenario. The goal was to design and tune a recommendation system able to predict past users' interactions, for the offline stage, and to provide recommendations pushed every day to real users through the XING portal, for the online stage. Our strategy, which saw models coming from different techniques combined in a multi-layer ensemble, granted us the first place in the offline part and the qualification as second best team in the final leaderboard. All our algorithms mainly resort to content-based approaches, that, thanks to its ability to provide good recommendations even for cold-start items allowed us, quite unexpectedly, to achieve good results in terms of prediction quality and computational time.
- Qilong Ba, Xiaoyong Li, and Zhongying Bai. 2013. Clustering collaborative filtering recommendation system based on SVD algorithm. In ICSESS, 2013 4th IEEE International Conference on. IEEE, 963--967.Google Scholar
- Ricardo Baeza-Yates, Berthier Ribeiro-Neto, et al. 1999. Modern information retrieval. Vol. 463. ACM press New York. Google ScholarDigital Library
- Leonardo Cella, Stefano Cereda, Massimo Quadrana, and Paolo Cremonesi. 2017. Derive item features relevance from past user interactions. In UMAP. Google ScholarDigital Library
- Wei Chu and Seung-Taek Park. 2009. Personalized recommendation on dynamic content using predictive bilinear models. In Proceedings of the 18th International Conference on World Wide Web. ACM, 691--700. Google ScholarDigital Library
- Marco de Gemmis, Pasquale Lops, Cataldo Musto, Fedelucio Narducci, and Giovanni Semeraro. 2015. Semantics-aware content-based recommender systems. In Recommender Systems Handbook. Springer, 119--159. Google ScholarDigital Library
- Asmaa Elbadrawy and George Karypis. 2013. Feature-based similarity models for top-n recommendation of new items. Department of Computer Science, University of Minnesota, Minneapolis, Minnesota, Tech. Rep (2013), 14--016.Google Scholar
- Abel Fabian, Yashar Deldjoo, Mehdi Elahi, and Daniel Kohlsdorf. 2017. RecSys Challenge 2017: Offline and Online Evaluation. In Proceedings of the 11th ACM Conference on Recommender Systems (RecSys '17). ACM, Como, ITALY, 2. Google ScholarDigital Library
- Zeno Gantner, Lucas Drumond, Christoph Freudenthaler, Steffen Rendle, and Lars Schmidt-Thieme. 2010. Learning attribute-to-feature mappings for coldstart recommendations. In Data Mining (ICDM), 2010 IEEE 10th International Conference on. IEEE, 176--185. Google ScholarDigital Library
- Yifan Hu, Yehuda Koren, and Chris Volinsky. 2008. Collaborative filtering for implicit feedback datasets. In Data Mining (ICDM), 2008 8th IEEE International Conference on. IEEE, 263--272. Google ScholarDigital Library
- Yehuda Koren and Robert Bell. 2015. Advances in collaborative filtering. In Recommender systems handbook. Springer, 77--118. Google ScholarDigital Library
- Melanie Mitchell. 1998. An introduction to genetic algorithms. MIT press. Google ScholarDigital Library
- Michael JD Powell. 1978. A fast algorithm for nonlinearly constrained optimization calculations. In Numerical analysis. Springer, 144--157.Google Scholar
- Andrew I Schein, Alexandrin Popescul, Lyle H Ungar, and David M Pennock. 2002. Methods and metrics for cold-start recommendations. In Proceedings of the 25th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. ACM, 253--260. Google ScholarDigital Library
- Mohit Sharma, Jiayu Zhou, Junling Hu, and George Karypis. 2015. Feature-based factorized bilinear similarity model for cold-start top-n item recommendation. In SIAM International Conference on Data Mining. SIAM, 190--198.Google ScholarCross Ref
Index Terms
- Content-Based approaches for Cold-Start Job Recommendations
Recommendations
Practical Lessons for Job Recommendations in the Cold-Start Scenario
RecSys Challenge '17: Proceedings of the Recommender Systems Challenge 2017The 2017 ACM RecSys Challenge focuses on the problem of job recommendations on XING in a cold-start scenario. In this paper we describe our solution as well as some practical lessons learned from the competition. We model this task as a binary ...
Naïve filterbots for robust cold-start recommendations
KDD '06: Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data miningThe goal of a recommender system is to suggest items of interest to a user based on historical behavior of a community of users. Given detailed enough history, item-based collaborative filtering (CF) often performs as well or better than almost any ...
Expediting Exploration by Attribute-to-Feature Mapping for Cold-Start Recommendations
RecSys '17: Proceedings of the Eleventh ACM Conference on Recommender SystemsThe item cold-start problem is inherent to collaborative filtering (CF) recommenders where items and users are represented by vectors in a latent space. It emerges since CF recommenders rely solely on historical user interactions to characterize their ...
Comments