search

SalesforceAIResearch / uni2ts

[ICML2024] Unified Training of Universal Time Series Forecasting Transformers
Apache License 2.0
612 stars 49 forks source link

Exogenous #35

Closed chenghaoliu89 closed 1 month ago

chenghaoliu89 commented 1 month ago

include the example of forecasting with exogenous variable

Closes #24, #33

gorold commented 1 month ago

Could we just use the existing dataframe which has multiple time series (a, b, c, ...) rather than creating another function for the synthetic data?

Also, there should be some header between the long/wide dataframe portions, or just include either one.

gorold commented 1 month ago

Could we add a brief explanation of what is the difference between past_feat_dynamic_real and feat_dynamic_real? Other than that, lgtm.

chenghaoliu89 commented 1 month ago

Could we add a brief explanation of what is the difference between past_feat_dynamic_real and feat_dynamic_real? Other than that, lgtm.

Yes, @gorold . it adds plot and description for each cases. It also add an example without using gluonts evaluation