Closed Gpen997 closed 3 years ago
https://github.com/winedarksea/AutoTS/blob/master/extended_tutorial.md
If you could be more specific, that would be helpful
I want to predict confidence intervals for data 0.95 and 0.90 How to split data for test and train in autots
If you read the documentation and the docstrings you will see answers to this...
For prediction interval, pass prediction_interval=0.95
(or whatever you desire) to the AutoTS() init. Then access the results:
prediction = model.predict()
upper_forecasts_df = prediction.upper_forecast
lower_forecasts_df = prediction.lower_forecast
and for train/test split, this is also clearly discussed in the documentation.
model = AutoTS(... other parameters...
num_validations=2,
validation_method='even',
)
Read the extended_tutorial for a discussion of different train/test split options.
How to predict time intervals in auto_ts in python