I have censored survival data. The Gaussian distribution appears to be an excellent fit. How can I compute an R-square statistic? The model has no covariates; it only estimated mu and sigma.
You may want to have a look on the following two papers: Korn and Simon (1990) give a theoretical background and some critical remarks on measures of explained variance, and Patrick Royston describes the Stata programs str2ph and str2d to compute R² statistics in survival models.
I had a look at the Korn paper. After giving many different metrics, they say that squared error loss should be used, but do not make clear what the formula is. The formula (1) assumes the existence of covariates. Do you know an R package that will compute R2?
The function OXS of the 'survAUC' package might do the trick. It is referring to the coefficient proposed by O'Quigley and collegues (2005). Other measures are included as well.