This message was deleted.
# general
s
This message was deleted.
j
Hey. I think you could run MLorecast.ts.predict passing a function that selects the features your model uses through before_predict_callback, something like:
Copy code
from functools import partial

def select_features(df, features):
    return df[features]

features_model1 = partial(select_features, features=['x1, 'x2'])
features_model2 = partial(select_features, features=['x2', 'x3'])
predictions = []
for model, callback in zip([model1, model2], [features_model1, features_model2]):
    model_preds = fcst.ts.predict({'model': model}, before_predict_callback=callback)
    predictions.append(model_preds)
e
yep this seems to work! Thank you!