[SciPy-user] Bias in numpy.random.multivariate_normal?

John Reid j.reid@mail.cryst.bbk.ac...
Sun Nov 11 04:32:44 CST 2007


Bryan Nollett wrote:
> The covariance matrix provided in sigma is not symmetric.

Good point. I needed some test covariance matrices and I completely 
forgot the symmetric requirement. There is no way to draw from the 
Wishart or Inv-Wishart distributions is there? Their pdfs would be handy 
too.

Thanks,
John.



More information about the SciPy-user mailing list