PMML Standard

A PMML Standard is a XML-based Markup Language Standard to represent prediction models as file-based prediction structure.

• AKA: Predictive Model Markup Language.
• Context:
  • It can represent a Decision Tree-based Prediction Model. http://www.dmg.org/v4-0-1/TreeModel.html
  • It can represent a Cluster Prediction Model. http://www.dmg.org/v4-0-1/ClusteringModel.html
  • It can represent an Association Rules-based Prediction Model. http://www.dmg.org/v4-0-1/AssociationRules.html
  • It can represent a Naive Bayes-based Prediction Model. http://www.dmg.org/v4-0-1/NaiveBayes.html
  • It cannot (yet) represent a Regression Tree-based Prediction Model or a Ranking Tree-based Prediction Model.
  • It cannot (yet) represent a Sequential Tagging Model, such as a Hidden Markov-based Model or a CRF-based Model.
• Example(s):
  • a PMML v4.0 Standard, http://www.dmg.org/pmml-v4-0-1.html
    • PMML v4.3(2016-08-02)[1]
  • a PMML v3.1 Standard, http://www.dmg.org/pmml-v3-1.html
  • a PMML v2.1 Standard, http://www.dmg.org/pmml-v2-1.html
• Counter-Example(s):
  • Gluon-API.
• See: XML Standard, XML Schema, DMG Group, PMML-Compliant Application.

---

---

References

2014

• (Wikipedia, 2014) ⇒ http://en.wikipedia.org/wiki/Predictive_Model_Markup_Language Retrieved:2014-8-12.
  • The Predictive Model Markup Language (PMML) is an XML-based file format developed by the Data Mining Group to provide a way for applications to describe and exchange models produced by data mining and machine learning algorithms. It supports common models such as logistic regression and feedforward neural networks.

    Since PMML is an XML-based standard, the specification comes in the form of an XML schema.

---