Censored - Survival Analysis in Tidymodels

on YouTube ↗18m

" summary

censored is a parsnip extension that adds a new model type proportional hazards and a new mode censored regression, plus engines for survival, glmnet and tree-based backends so you can swap models without reworking data. It keeps a formula interface for glmnet, returns predictions as tibbles and introduces predict type: time and survival (list-column of probabilities at requested times). Works with workflows with a small formula workaround; examples live at the repo.

AI-written summary · about

" chapters (5)

1. 0:00Introduction
2. 1:44Survival Analysis
3. 4:10Specify and fit models
4. 10:28Predicting
5. 16:01Questions

" related talks

• 18mCensored - Survival Analysis in TidymodelsHannah Frick · rstudio::conf(2022)
• 19mtidymodels for time-to-event dataHannah Frick · posit::conf(2024)
• 20mEvaluating Time-to-Event Models is HardMax Kuhn · posit::conf(2024)
• 01h 24mBayesianism and Survival AnalysisJake Coltman, Jacob Goodwin · PyData London 2016
• 18mtidymodels: Adventures in Rewriting a Modeling PipelineRyan Timpe · posit::conf(2023)
• 15mThe tidysynthesis R packageAaron R. Williams · rstudio::conf(2022)
• 34mFast, beautiful and easy Bayesian modelling: Can you have it all?Rhydwyn McGuire · PyCon Australia 2015
• 01h 23mPredictive survival analysis with scikit-learn, scikit-survival and lifelinesOlivier Grisel · PyData Global 2023