TOWARDS ONLINE COMPOSITION OF PMML PREDICTION MODELS

Diana Gorea

Abstract

The paper presents a general context in which composition of prediction models can be achieved within the boundaries of an online scoring system called DeVisa. The system provides its functionality via web services and stores the prediction models represented in PMML in a native XML database. A language called PMQL is defined, whose purpose is to process the PMML models and to express consumers’ goals and the answers to the goals. The composition of prediction models can occur either implicitly within the process of online scoring, or explicitly, in which the consumer builds or trains a new model based on the existing ones in the DeVisa repository. The main scenarios that involve composition are adapted to the types of composition allowed in the PMML specification, i.e sequencing and selection.

References

  1. Axis (2007). Apache axis. http://ws.apache.org/axis/.
  2. DeVisa (2007). Devisa. http://devisa.sourceforge.net.
  3. Fensel, D., Lausen, H., Polleres, A., de Bruijn, J., Stollberg, M., Roman, D., and Domingue, J. (2007). Enabling Semantic Web Services. Springer, Berlin Heidelberg.
  4. Gorea, D. (2008). Devisa concepts and architecture of a data mining models scoring and management web system. In Proceedings of the 10th International Conference on Enterprise Information Systems. INSTICC.
  5. Ian H. Witten, E. F. (2005). Data Mining Practical Machine Learning Tools and Techniques. Morgan Kaufmann series in data management systems. Elsevier.
  6. Kuncheva, L. I. (2004). Classifier ensembles for changing environments. In Proc. 5th Int. Workshop on Multiple Classifier Systems, volume 3077 of LNCS, pages 1- 15. Springer-Verlag.
  7. M.Harries, Sammut, C., and Horn, K. (1998). Extracting hidden context. Machine Learning, 36(2):101-126.
  8. PMML (2007). Pmml http://www.dmg.org/pmml-v3-2.html.
  9. Prodromidis, A., Chan, P., and Stolfo, S. (2000). Metalearning in distributed data mining systems: Issues and approaches, chapter 3. AAAI/MIT Press.
Download


Paper Citation


in Harvard Style

Gorea D. (2008). TOWARDS ONLINE COMPOSITION OF PMML PREDICTION MODELS . In Proceedings of the Third International Conference on Software and Data Technologies - Volume 3: ICSOFT, ISBN 978-989-8111-53-1, pages 300-303. DOI: 10.5220/0001890503000303


in Bibtex Style

@conference{icsoft08,
author={Diana Gorea},
title={TOWARDS ONLINE COMPOSITION OF PMML PREDICTION MODELS},
booktitle={Proceedings of the Third International Conference on Software and Data Technologies - Volume 3: ICSOFT,},
year={2008},
pages={300-303},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001890503000303},
isbn={978-989-8111-53-1},
}


in EndNote Style

TY - CONF
JO - Proceedings of the Third International Conference on Software and Data Technologies - Volume 3: ICSOFT,
TI - TOWARDS ONLINE COMPOSITION OF PMML PREDICTION MODELS
SN - 978-989-8111-53-1
AU - Gorea D.
PY - 2008
SP - 300
EP - 303
DO - 10.5220/0001890503000303