search

Nixtla / nixtla

TimeGPT-1: production ready pre-trained Time Series Foundation Model for forecasting and anomaly detection. Generative pretrained transformer for time series trained on over 100B data points. It's capable of accurately predicting various domains such as retail, electricity, finance, and IoT with just a few lines of code 🚀.
https://docs.nixtla.io
Other
1.91k stars 151 forks source link

Incorporating Exogenous Variables #266

Closed drRussClay closed 1 month ago

drRussClay commented 3 months ago

Hello - I've been following along with the installation guide for TimeGPT in Python. I want to make sure I understand how to add exogenous variables to the forecast model. It seems that I can only pass an X_df that covers the forecast period, correct?

In my current test problem, I am adding a single exogenous variable that is very highly correlated with the outcome that I am forecasting throughout the dataset. There is one exception, where the relation gets decoupled, and that is during the forecast period. I expected to see the forecast follow the trend of exogenous predictor, but it does not - it remains smooth despite the sudden change in the exogenous variable, which makes me wonder if I am adding it correctly to the forecast.

Can you confirm that X_df should only contain data for the forecast period? I get errors if try to pass an X_df with any other length.

Thank you!

jmoralez commented 2 months ago

Hey @drRussClay, thanks for using TimeGPT. Yes, X_df should contain the values of the exogenous features during the forecast period. Can you provide a small example with fake data that reproduces what you're seeing?

github-actions[bot] commented 1 month ago

This issue has been automatically closed because it has been awaiting a response for too long. When you have time to to work with the maintainers to resolve this issue, please post a new comment and it will be re-opened. If the issue has been locked for editing by the time you return to it, please open a new issue and reference this one.