Modeling Dynamical Systems with Data Stream Mining

Aljaž Osojnik1,2, Panče Panov1 and Sašo Džeroski1,2,3

  1. Jožef Stefan Institute
    Jamova cesta 39, Ljubljana, Slovenia
    faljaz.osojnik, pance.panov, saso.dzeroskig@ijs.si
  2. Jožef Stefan Interntaional Postgraduate School
    Jamova cesta 39, Ljubljana, Slovenia
  3. CIPKeBiP
    Jamova cesta 39, Ljubljana, Slovenia

Abstract

We address the task of modeling dynamical systems in discrete time using regression trees, model trees and option trees for on-line regression. Some challenges that modeling dynamical systems pose to data mining approaches are described: these motivate the use of methods for mining data streams. The algorithm FIMT-DD for mining data streams with regression or model trees is described, as well as the FIMT-DD based algorithm ORTO, which learns option trees for regression. These methods are then compared on several case studies, i.e., tasks of learning models of dynamical systems from observed data. The experimental setup, including the datasets, and the experimental results are presented in detail. These demonstrate that option trees for regression work best among the considered approaches for learning models of dynamical systems from streaming data.

Key words

dynamical systems, data streams, data mining, regression and model trees, option trees

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS150518009O

Publication information

Volume 13, Issue 2 (June 2016)
Year of Publication: 2016
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium

Full text

DownloadAvailable in PDF
Portable Document Format

How to cite

Osojnik, A., Panov, P., Džeroski, S.: Modeling Dynamical Systems with Data Stream Mining. Computer Science and Information Systems, Vol. 13, No. 2, 453–473. (2016), https://doi.org/10.2298/CSIS150518009O