Gabriel Gimenez

This is not intended behavior but there isn't any parameter that will improve the scoring functions or any simple fix. We'll add the refactor and optimization of the scoring methods...

Hi Jayant, We currently don't have any way to deal with time varying features, but it's a really interesting problem. We'll keep an eye on it and see if we...

You can follow lifelines recommendations, except for `3. Is there high-collinearity in the dataset? Try using the variance inflation factor (VIF) to find redundant variables.` as it shouldn't make a...

We usually use `joblib` `dump` and `load` to save and load models. It's important to fix xgboost, sklearn and lifelines versions when loading, to avoid some difference in the underlying...

Hi @esbraun! We actually don't have a way to get a survival at a specific time, I'll add it to future improvements. Using Kaplan neighbours or the StackedWeibull method it's...

@igorkf This warning occurs when a time bucket you don't have any event or only events, so that specific logistic regression for that time bucket will only predict 0 or...