What does the predictions contain ?

[Clément Rieux]

Hi !

I ran the notebook and was wondering what exactly contain predictions ?

predictions = Exp.predict(setting=setting, pred_data=test_data)

We have indeed the projection of 96 and the 7 variables columns but I don't understand the dimension.

Thanks

[Xu-Yang Chen]

The shape of the prediction is (11521, 96, 7), in which 11521 is the number of samples, 96 is the prediction length, and 7 is the number of channels.

[Clément Rieux]

For example, i used predictions_original_scale = scaler.inverse_transform(predictions[1]) predictions_df = pd.DataFrame(predictions_original_scale, columns=test_data.columns[1:]) to get data with original value.

Is it correct if I say that I have this datframe contains the predictions values ​​of the next 96 periods?
And in this case predictions_original_scale = scaler.inverse_transform(predictions[2] contains +1 period ?

My question is how do I predict in the future in one time ?