JoaquinAmatRodrigo / skforecast

Time series forecasting with machine learning models
https://skforecast.org
BSD 3-Clause "New" or "Revised" License
996 stars 113 forks source link

add method to forecasters to return the input data that is passed to the model to make predictions #647

Open solegalli opened 4 months ago

solegalli commented 4 months ago

Currently, there is no way to see which data is being used to make predictions.

create_X_y creates the data that is used to train the model, but it;s not able to create data that is passed to "predict()`.

The data for the prediction is created in predict() but it is not exposed.

Could we capture the lines that create the input data to the regressor.predict() in a new method, say create_forecast_input() to expose that value to the user?

It'd help with debugging and understanding what the forecaster does under the hood

JoaquinAmatRodrigo commented 3 months ago

I agree. I like the idea that users can explore as much as possible all the work that skforecast does internally. I have added it to the backlog for the next release.

JavierEscobarOrtiz commented 1 week ago

Hello @solegalli,

The create_predict_inputs method will be available in skforecast 0.13.0. 😄