Predictive Model Markup Language (PMML) is an XML standard for the representation and exchange of data mining models.
PMML is developed and owned by the Data Mining Group, an “independent, vendor led consortium that develops data mining standards” which appears to be led by IBM. All the full members (and most of the others) are from the USA.
The full PMML standard is available on their website with an open licence (do what you like with it, but quote this licence and dont blame us if it goes wrong) and some conversion/ validation tools.
DMG claim “With PMML, it is straightforward to develop a model on one system using one application and deploy the model on another system using another application”.
DMG give examples of models built using PMML; such as SAS Enterprise Miner which claims: “Forward-thinking organizations today are using SAS data mining software to detect fraud, minimize credit risk, anticipate resource demands, increase response rates for marketing campaigns and curb customer attrition.”
As Im reading the Learning XML at the moment, with a vague hope of one day finding a really good way I can use this verbose and complex (but brilliant) technology, the idea of PMML appeals. For example, could it be used to make everyones models of the spread of flu compatible, so that at last we can have simulations which can run with each others data or assumptions?