I am interested in using survival models together with the tidymodels package. I have come across the 'censored' package, which works in conjunction with tidymodels. However, I need to tune the model parameters and use it together with the 'recipes' package. Additionally, I believe that 'censored' does not currently support XGBoost models. Is there a plan to implement parameter tuning in the 'censored' package?
We have tuning and several other aspects of modeling event time data implemented. We are currently doing a lot of testing and documentation right now. The code is mostly in the Github versions of several packages.
There are also two articles (here and here) on tidymodels.org that talk about evaluating models.
I hope to publish a blog post before we go to the posit conference in September that has examples, installation instructions, and a request for people to test it out.
For xgboost, there is a pull requisition in censored. As with all things xgboost, they go about structuring the data in the weirdest way so that needs a lot of work to implement (we might just rewrite it).
Thank you, Max, for the prompt response.
I have been working with a model to predict the kidney transplant waiting list in São Paulo, Brazil. In this model, I used Cox regression. (A machine learning prediction model for waiting time to kidney transplant)
Currently, I am in the process of predicting kidney transplants across all Brazilian states. To achieve this goal, I intend to employ tidymodels and the censored package. I am hopeful that these implementations will be operational soon.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.
If you have a query related to it or one of the replies, start a new topic and refer back with a link.