Standards for representing uncertainty

From Intamap

The results of INTAMAP will always include some information on (posterior) uncertainty. This must be communicated to the user in a standard manner. There are no current standards I am aware of. Some relevant ISO standards are 19113/4/5, Quality, Quality and meta-information. The ISO more or less says: (i) you can report quality measures the way you want, (ii) this information should go into the meta-information. The writers have left some freedom how to express quality, but I believe they include distribution functions for continuous variables. While we may need to extend existing standards, our rationale should be based on the fact that these ISO standards are not enough for us, or nobody will buy our proposal. {Edzer}

It was never my intention to try to force a new standard where there is currently one in place. ISO is important and I am very glad you have followed this up. I should 'own up' that I have an agenda here that extends beyond INTAMAP (to a project called MUCM I am working with, and also to a failed, but still possible, project I wanted to do) which is about the interoperability of statistical models; I see the uncertainty schema as providing a prototype for this. I have done some, but not extensive, research and have found very little. ISO was something I had not yet checked, so it is very good to have some confirmation. I might have to look at some other ISO standards however ... to be done! {Dan}