You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Long/Short Term Forecasting, Imputation, Classification and Anomaly Detection is a great set of task based categorical features to test for all the models.
But why is Multivariate Regression missing from this "comprehensive benchmark"?
I would suggest, for the benchmark to be truly comprehensive, the regression task should be part of the benchmark, if this is to be used as a way of ranking and as a standard quality measure for all time series models in the future.
The text was updated successfully, but these errors were encountered:
Hi, many thanks for your suggestion.
'Multivariate Regression' is important. But to our best knowledge, this task is not well-studies in recent papers of machine learning community.
Would you please kindly give us some related works or benchmarks?
Long/Short Term Forecasting, Imputation, Classification and Anomaly Detection is a great set of task based categorical features to test for all the models.
But why is Multivariate Regression missing from this "comprehensive benchmark"?
I would suggest, for the benchmark to be truly comprehensive, the regression task should be part of the benchmark, if this is to be used as a way of ranking and as a standard quality measure for all time series models in the future.
The text was updated successfully, but these errors were encountered: