There are several variants of pseudo \(R^2\) that can be computed for a likelihood estimation. They all relate the log likelihood of the estimated model to the log likelihood of the null model.
The ones included here are McFadden's, Adjusted McFadden's, Cox & Snell's, and Nagelkerke, Cragg, and Uhler's.
pseudoR2(opt)returned value from mphcrm.
A matrix is returned, with one row for each iteration containing the various pseudo \(R^2\)s.