{"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/effective-ways-to-build-and-evaluate","title":"Effective Ways to Build and Evaluate Individual Survival Distributions","arxiv_id":"1811.11347","date":"2018-11-28","proceeding":null,"authors":["Humza Haider","Bret Hoehn","Sarah Davis","Russell Greiner"],"abstract":"An accurate model of a patient's individual survival distribution can help\ndetermine the appropriate treatment for terminal patients. Unfortunately, risk\nscores (e.g., from Cox Proportional Hazard models) do not provide survival\nprobabilities, single-time probability models (e.g., the Gail model, predicting\n5 year probability) only provide for a single time point, and standard\nKaplan-Meier survival curves provide only population averages for a large class\nof patients meaning they are not specific to individual patients. This\nmotivates an alternative class of tools that can learn a model which provides\nan individual survival distribution which gives survival probabilities across\nall times - such as extensions to the Cox model, Accelerated Failure Time, an\nextension to Random Survival Forests, and Multi-Task Logistic Regression. This\npaper first motivates such \"individual survival distribution\" (ISD) models, and\nexplains how they differ from standard models. It then discusses ways to\nevaluate such models - namely Concordance, 1-Calibration, Brier score, and\nvarious versions of L1-loss - and then motivates and defines a novel approach\n\"D-Calibration\", which determines whether a model's probability estimates are\nmeaningful. We also discuss how these measures differ, and use them to evaluate\nseveral ISD prediction tools, over a range of survival datasets.","url_abs":"http://arxiv.org/abs/1811.11347v1","url_pdf":"http://arxiv.org/pdf/1811.11347v1.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":"effective-ways-to-build-and-evaluate","repo_url":"https://github.com/haiderstats/ISDEvaluation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"effective-ways-to-build-and-evaluate","repo_url":"https://github.com/loft-br/xgboost-survival-embeddings","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[],"methods":[{"method_slug":"logistic-regression","method_name":"Logistic Regression"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1811.11347","atlas_url":"https://app.syntology.ai/?focus=1811.11347","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}