KimMeen/Time-LLM

Long-term Forecasting Target

Closed this issue · 1 comments

Thanks for your great work!

In the ETT datasets, the target shoule be the oil temperature (OL), but in the repo I find that Time-LLM seems to predict all 7 features in a single-to-single manner. Also the feature category is not part of the input. So Time-LLM seems to hypothesize that all 7 features are from the same distribution. Is this the same setting with other baseline methods? Your paper claim that OL is the only target feature.

Hi, currently the ETT dataset is primarily a multivariate prediction dataset, which predicts all 7 values including oil temperature (OL). Additionally, for the sake of operational efficiency, we have employed channel-independent techniques in the dataloader to split it into univariate predictions for processing.