esbraun

Thank you! Is this something that you're planning on merging eventually as well?

@strakehyr I've run into issues using the train() method with covariates as well. I think it's because the TFT decoder expects future observations for future known covariates (vs ones known...

I agree it's important for econml to accept missing values to support algorithms that directly handle missing values (i.e. most notably xgboost, lightgbm and catboost). Forcing the imputation of missing...

Update: I've empirically found significant improvement using the negative binomial distribution for a operations related count valued forecasting problem using the TFT implementation from kashif's repo over the mxnet TFT...

Seconding @olamagnusandersson. I’d still rather see a warning than an error thrown for the reasons below. > I agree it's important for econml to accept missing values to support algorithms...

I've also noticed TFT doesn't play well with PandasDataset dynamic real features. I think it's because TFT expects the resulting ListDataset to include the features by name, not grouped into...

@erasedcitizen11 I created a fork that disables the missing value checks which is a requirement for our use cases. Use at your own risk: https://github.com/esbraun/EconML/tree/main

One possibility is to train two different DynamicDML learners: one for treatment A and the other for treatment B. For the treatment A learner, include features in X that describe...