CosmoMC: 2D marginalized plots

Use of Healpix, camb, CLASS, cosmomc, compilers, etc.
Post Reply
Jaiseung Kim
Posts: 13
Joined: January 21 2008
Affiliation: Max Planck Institute for Astrophysics

CosmoMC: 2D marginalized plots

Post by Jaiseung Kim » January 28 2008

I have run CosmoMC with a new parameter by re-interpreting \Omega_k.
2D marginalized plots shows the colored 2D images with two contours.
What is puzzling is the two contours do not match color-fill well.
(For instance, the highest point (by color) of some plot is outside 1-sigma contour.)
Is it acceptable? If so, is it because they represent numbers from different statistics like a dotted and a solid line in 1D plot?

thank you,

Antony Lewis
Posts: 1594
Joined: September 23 2004
Affiliation: University of Sussex

Re: CosmoMC: 2D marginalized plots

Post by Antony Lewis » January 28 2008

Yes, shading is by mean likelihood like dotted lines in 1D. So it can be right if the distribution is non-Gaussian.

Post Reply