{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/fifty-shades-of-ratings-how-to-benefit-from-a","title":"Fifty Shades of Ratings: How to Benefit from a Negative Feedback in Top-N Recommendations Tasks","arxiv_id":"1607.04228","date":"2016-07-14","proceeding":null,"authors":["Evgeny Frolov","Ivan Oseledets"],"abstract":"Conventional collaborative filtering techniques treat a top-n recommendations\nproblem as a task of generating a list of the most relevant items. This\nformulation, however, disregards an opposite - avoiding recommendations with\ncompletely irrelevant items. Due to that bias, standard algorithms, as well as\ncommonly used evaluation metrics, become insensitive to negative feedback. In\norder to resolve this problem we propose to treat user feedback as a\ncategorical variable and model it with users and items in a ternary way. We\nemploy a third-order tensor factorization technique and implement a higher\norder folding-in method to support online recommendations. The method is\nequally sensitive to entire spectrum of user ratings and is able to accurately\npredict relevant items even from a negative only feedback. Our method may\npartially eliminate the need for complicated rating elicitation process as it\nprovides means for personalized recommendations from the very beginning of an\ninteraction with a recommender system. We also propose a modification of\nstandard metrics which helps to reveal unwanted biases and account for\nsensitivity to a negative feedback. Our model achieves state-of-the-art quality\nin standard recommendation tasks while significantly outperforming other\nmethods in the cold-start \"no-positive-feedback\" scenarios.","url_abs":"http://arxiv.org/abs/1607.04228v1","url_pdf":"http://arxiv.org/pdf/1607.04228v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"fifty-shades-of-ratings-how-to-benefit-from-a","repo_url":"https://github.com/Evfro/fifty-shades","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"fifty-shades-of-ratings-how-to-benefit-from-a","repo_url":"https://github.com/Evfro/polara","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"collaborative-filtering","task_name":"Collaborative Filtering"},{"task_slug":"recommendation-systems","task_name":"Recommendation Systems"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}