An example on heteroskedastic noise with StudentT likelihood #1730
-
Is there a feature that needs some example code in a notebook? Hi, but with Student-t likelihood as described in http://proceedings.mlr.press/v51/saul16.html. I can not just change this:
into this:
because then I get the following error in the optimization: An example on this would be helpful. Kind regards, Ivan |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
This happens because There are some possibilities to work around it. If you want a fixed value for the
If you want to have the |
Beta Was this translation helpful? Give feedback.
-
Great, I thank you very much! |
Beta Was this translation helpful? Give feedback.
This happens because
HeteroskedasticTFPConditional
calls thedistribution_class
asdistribution_class(loc, scale)
, using positional arguments. However,tfp.distributions.StudentT
expects signature istfp.distributions.StudentT(df, loc, scale)
.There are some possibilities to work around it. If you want a fixed value for the
df
parameter, you can just call:If you want to have the
df
parameter not fixed, I'd suggest subclassingMultiLatentTFPConditional
.