Abnormal normals

I’m doing thesis revisions at the moment and one of the examiners has asked that I refit some of the models in one of the chapters. The chapter corresponds to a book chapter that I wrote with one of my supervisors where we looked at Bayesian spline regression. This was pretty early in my PhD and represented a stepping stone from the frequentist approach that Simon Wood uses in his mgcv pacakge (used for my first paper) and the Bayesian approach to Generalised Additive Models with autoregressive errors (based on Siddartha Chib’s 1993 paper).

One thing that I’ve discovered that really bugs me in MATLAB is that normpdf (density function of a normal distribution) is specified in terms of a mean and standard deviation whereas mvnpdf (density function of a multivariate normal) is specified in terms of a mean vector and covariance matrix. The result is that normpdf(0,0,1.5) and mvnpdf(0,0,1.5) give different results! In my mind, the univariate normal is a special case of the multivariate normal and the usage should be consistent.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s