search

HyunWookL / TESTAM

Official Code of TESTAM: A Time-Enhanced Spatio-Temporal Attention Model with Mixture of Experts
MIT License
18 stars 1 forks source link

If the baseline results reported in the paper are inconsistent with the corresponding paper. #1

Open ownnaruto opened 3 months ago

ownnaruto commented 3 months ago

The 'GTS' results reported in your paper are inconsistent with the corresponding paper, as detailed below:

you paper：

截屏2024-03-20 下午4 45 55

original paper：

截屏2024-03-20 下午4 46 49

How this happen？

HyunWookL commented 3 months ago

Thank you for your comment! Actually, the same problem occurred in the very recent article on traffic forecasting, MegaCRN. Please refer to the MegaCRN paper, AAAI'23. In this paper (at the same time, our paper), we have noticed that some models have a reproducibility issue, so we updated our results with our experiment. I hope this comment helps you understand such inconsistency.