Survival analysis is now supported across the tidymodels framework, a collection of R packages for modeling and machine learning using tidyverse principles. It covers the entire predictive modeling workflow from data splitting, resampling, feature engineering, model fitting, and performance evaluation to tuning. It provides a consistent interface with composable functions that allow beginners a safe start and advanced users access to more specialized techniques such as feature engineering on text data or tuning via racing methods.
The recent addition of dedicated performance metrics has enabled us to support tuning of survival models and unlock the entire framework for survival analysis. This workshop focuses on the core components of tidymodels to get you up and running with predictive survival analysis.
This workshop is for you if you:
- are familiar with basic survival analysis such as censoring of time-to-event data, Kaplan-Meier curves, proportional hazards models
- are familiar with the basic predictive modeling workflow such as split in train and test set, resampling, tuning via grid search
- want to learn how to leverage the tidymodels framework for survival analysis
PreparationPlease see the **Preparation** section on
https://hfrick.github.io/tidymodels-survival-workshop/ prior to the day for what to install.