User:Icemanmark

Well, i'll start with a hello!

I created a Wikipedia profile to update on a small error on the Generalized linear model page under Link Functions. I'm busy studying Statistics and Data Science at MITx and discovered the error in association with the Canonical Link function $$g(\mu)$$for the Binomial Distribution. I noticed this while solving a homework problem in class, i cross referenced and found that the Wikipedia reference did not follow the same result. My reason for the change is when representing the link function $$g(\mu)$$ with $$\mu$$ and not $$p$$ itself the canonical link function is represented with an $$n$$ instead of a $$1$$. A clear explanation of the the derivation is shown on Stack Exchange Cross Validated.

The change i made was from this $$\mathbf{X}\boldsymbol{\beta}=\ln \left(\frac \mu {1-\mu}\right) \,\!$$   to this   $$\mathbf{X}\boldsymbol{\beta}=\ln \left(\frac \mu {n-\mu}\right) \,\!$$.

Thanks to Prof Rigollet, Dr. Yaroslav, Dr. Karene and the rest of the course team for a great course.