GAUSSIAN PROCESS REGRESSION (GPR) REPRESENTATION IN PREDICTIVE MODEL MARKUP LANGUAGE (PMML)

Published

March 29, 2017

Author(s)

Jinkyoo Park, David Lechevalier, Ronay Ak, Max K. Ferguson, Kincho H. Law, Yung-Tsun Lee, Sudarsan Rachuri

Abstract

This paper describes a Predictive Model Markup Language (PMML) representation of Gaussian Process Regression (GPR) models. PMML is an XML-based standard language used to represent data mining and predictive analytic models, as well as pre- and post-processed data. Although the current PMML version, PMML 4.2, covers several data mining algorithms, it does not provide capabilities for representing probabilistic (stochastic) machine learning algorithms such as GPR, Bayesian Networks and Probabilistic Support Vector Machine that are widely used for constructing predictive models with the uncertainty quantification capability. This paper focuses on the GPR algorithm and discusses its representation using PMML. Furthermore, we present a prototype to generate GPR PMML representation using a real data set in the manufacturing domain.

Citation

The ASTM Journal of Smart and Sustainable Manufacturing

Volume

1

Issue

1

Pub Type

Journals

Download Paper

DOI Link

Keywords

PMML, Gaussian Process Regression, Predictive Analytics, Data Mining, Standards, XML

Citation

Park, J. , Lechevalier, D. , Ak, R. , Ferguson, M. , Law, K. , Lee, Y. and Rachuri, S. (2017), GAUSSIAN PROCESS REGRESSION (GPR) REPRESENTATION IN PREDICTIVE MODEL MARKUP LANGUAGE (PMML), The ASTM Journal of Smart and Sustainable Manufacturing, [online], https://doi.org/10.1520/SSMS20160008 (Accessed October 14, 2025)