Open
Description
Hi, thank you so much for your work. It inspires me a lot. As most of the LSTM use case is using the output of last time step and concatenate it with current features, which can lead to better performance. Especially in the infectious disease field, we concatenate the output of current time step (newly infected cases) with the covariates of the next time step. I want to ask in this case, how can I use timeshap? Thank you
Metadata
Metadata
Assignees
Labels
No labels