Time Varying Overdispersed Poisson Process

Hi Kyle

This should work

seasonality = pm.Normal('seasonality', mu=y, sigma=sigma, shape=len(df1))

though I’m not sure why it is required as usually the shapes will broadcast themselves fine.
Also I changed the sd into sigma as sd will be depreciated soon.

After that I noticed that the next lines fail. k is not defined and nor is df1.user_id.
In addition a and b are not used at all?

1 Like