DMG Updates Predictive Model Standard

The Data Mining Group, a consortium of companies and organizations developing standards for statistical and data mining models, has rolled out Version 4.0 of its Predictive Model Markup Language (PMML).

PMML is a standard for statistical and data mining models, allowing developers to create a model using one application, and then deploy that model using another system and a different application.

The upgrade provides support for multiple models, including segmented models and ensembles of models. In addition, it features improved support for pre-processing data, to help simplify deployment.

'PMML simplifies the deployment of analytic models,' says Robert Grossman, Chair of the Data Mining Group. 'With PMML, deploying a new model is as easy as reading an XML file.'

Members of the Data Mining Group include IBM, MicroStrategy, the National Center for Data Mining, the Open Data Group, SAS, SPSS and Zementis.

Founding member SPSS has incorporated the latest edition of PMML into its PASW Modeler (formerly Clementine) data mining workbench and PASW Statistics tools, to enable customers to apply modeling techniques and visualizations.

Web site: www.dmg.org .

All articles 2006-23 written and edited by Mel Crowther and/or Nick Thomas, 2024- by Nick Thomas, unless otherwise stated.