Paper Review: Temporal Fusion Forecasting (Christopher Ewanik)

[@saadbhai5998]

Thanks for the video.
Why TCN wasn't considered for performance comparison? 🤔

[@jonhuff9524]

You mention that the entire model up until the end is linear, ReLU for example introduces non-linearities, otherwise it just collapses to a linear model? Also can you elaborate on "non-linear data such as seasonal"? All time series are gonna have some component of non-linearity, more of a question of additive vs. multiplicative?

[@subbaramireddy-rp6uo]

Hi Can u share the link where we can access the ppt

[@samirelzein1095]

very interesting!