Tutorials on Cox and Poisson Regression

[Dr. V. Ravi]

Hi All:

Could anyone let me know the URLs for online tutorials on Cox and Poisson regression models, the underlying theory and method of fitting.

thanks in advance

[Pravin Kadekodi]

Even I have been trying to get some information on Cox-Snell residuals. Believe you are also looking for information on this.
Till now this is how I have understood the Cox-Snell residuals. The closer the points fit to the 45 degree line the better is the fit of the assumed distribution. This is a eye-ball method to estimate goodness of the fit.
One more parameter is the Cox-Snell residual value, the lower the value the better the fit, but I don't have more information on this..
If some one can provide better explaination that would help..

[DrDave]

There is a short description at http://www-personal.umich.edu/~bobwolfe/675/11Diagnostic/Diagnose.pdf

The key point is that if the assumed model is correct, the Cox-Snell residuals (which are not really residuals, but that's another post) should follow the exponential distribution. The diagnostic plot looks at an exponential probability plot of the CS residuals. If it appears plausibly exponential, the model is supported. Formal goodness of fit tests can be applied.

The (extremely terse) description on the Weibull site (at http://www.weibull.com/AccelTestWeb/cox_snell_residuals_plots.htm ) is good as far as it goes, with one quibble: I would show the Reliability function as estimated (with a 'hat') above it.

Hope this helps.