Question about evidence approximation when maximise wrt alpha (bishop chapter 3.5.2)

by Yong Zou   Last Updated August 14, 2019 00:19 AM

I'm reading 3.5.2 of PRML book and am confused about the result of taking derivative of marginal likelihood function with regard to alpha. In particular regarding this function:

enter image description here

And according to the book the derivative with regard to alpha is:

enter image description here

So it seems m_N is not a function of alpha at all. But clearly it is a function of alpha according to this formula:

enter image description here

enter image description here

I must have missed something very obvious.

Thanks!

Tags : regression


Related Questions


Multiple regression, full and restricted model

Updated March 12, 2017 19:19 PM

Models under Regression Analysis list

Updated October 15, 2018 00:19 AM

Fitting polynomial equation and combined effect

Updated November 14, 2018 13:19 PM

Restricted Weighted Linear Regression in R

Updated June 06, 2019 19:19 PM