Monthly Archives: May 2011

Stata-like Marginal Effects for Logit and Probit Models in R « TimeSeriesIreland

Stata-like Marginal Effects for Logit and Probit Models in R « TimeSeriesIreland. I thought this was rather interesting, but haven’t had time to think about it yet.

Posted in Uncategorized | Leave a comment

Which distribution fits

There’s a great blog article here: http://www.johndcook.com/blog/2010/08/11/what-distribution-does-my-data-have/ Nice discussion about choosing a distribution to fit to a set of data, but perhaps more importantly, a nice reminder that these are all abstractions. To reword de Finetti: There is no such … Continue reading

Posted in Uncategorized | Leave a comment

Distribution fitting

I’ve just dumped a load of rpanel animations here (all to do with distribution fitting for a Binomial).   (Need to tidy this up when my toothache gets better). The first bit of code just draws a picture of a standard … Continue reading

Posted in R, rpanel | Tagged , , , | 2 Comments

Why use Bayesian methods

(Here is an attempt to explain why Bayesian methods are the inference method of choice) O’Hagan (Kendall’s Library) suggests: Some data are observed, and we wish to make a statement (inference) about one or more unknown features of the system … Continue reading

Posted in Bayesian philosophy | Tagged , | Leave a comment