I am reading through the chapter and found some typos/inconsistencies - sorry for not making a pull request (and thanks for making this great book available!)
When introducing the sigmoid function you have
\text{logit}^{-1}(\xi) = \frac{e^\xi}{1 + e^{\xi}} = \frac{1}{1 + e^{-\xi}}
it should be
\text{logit}^{-1}(\xi) = \frac{e^\xi}{1 + e^{\xi}} = \frac{1}{1 + e^{-\xi}}-1
I believe.
Some lines above you have
(Note, many more machine learning focused texts will use p as the number of parameters. This is an arbitrary choice, but you should be aware of it.)
but you are also using p as the number of parameters, so I was confused by the sentence.
I am reading through the chapter and found some typos/inconsistencies - sorry for not making a pull request (and thanks for making this great book available!)
When introducing the sigmoid function you have
\text{logit}^{-1}(\xi) = \frac{e^\xi}{1 + e^{\xi}} = \frac{1}{1 + e^{-\xi}}
it should be
\text{logit}^{-1}(\xi) = \frac{e^\xi}{1 + e^{\xi}} = \frac{1}{1 + e^{-\xi}}-1
I believe.
Some lines above you have
(Note, many more machine learning focused texts will use p as the number of parameters. This is an arbitrary choice, but you should be aware of it.)
but you are also using p as the number of parameters, so I was confused by the sentence.