Abstract
A heuristic method of model choice for a nonlinear regression problem on real line, based on the Equation Finder (EF) of Zembowicz and Żytkow (1992), is proposed and discussed. In our implementations of the EF we use a new, actually a three-stage, procedure for stabilizing model selection. First, a set of pseudosamples is obtained from the original sample by resampling in some way. Second, for each pseudosample, a family of acceptable models is found by a clustering-like algorithm performed on models with largest (adjusted) coefficients of determination. And third, the final selection is made from among the models which appear most often in the families obtained in the second stage.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Breiman, L.: Heuristics of instability and stabilization in model selection. Ann. Statist. 24 (1996a) 2350–2393
Breiman, L.: Bagging predictors. Machine Learning 26 (1996b) 123–140
Friedman, J.H. and Stuetzle, W.: Smoothing of scatterplots. Technical Report, Stanford University, Dept. of Statist., 1982
Krzanowski, W.J.: Principles of multivariate analysis. Oxford University Press, 1988
Moulet, M.: Comparison of three inductive numerical law discovery systems. In: Machine learning and statistics, G. Nakhaeizadeh and C.C. Taylor (eds.), Wiley, 1997, 293–317
Tibshirani, R.: Estimating transformations for regression additivity and variance stabilization. J. American Statist. Assoc. 83 (1988) 394–405
Zembowicz, R. and Żytkow, J.M.: Discovery of equations: Experimental evaluation of convergence. In: Proceedings of the 10th National Conference on Artificial Intelligence, AAAI-92, AAAI Press, 1992, 70–75
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1998 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Ćwik, J., Koronacki, J. (1998). A Heuristic Method of Model Choice for Nonlinear Regression. In: Polkowski, L., Skowron, A. (eds) Rough Sets and Current Trends in Computing. RSCTC 1998. Lecture Notes in Computer Science(), vol 1424. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-69115-4_10
Download citation
DOI: https://doi.org/10.1007/3-540-69115-4_10
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-64655-6
Online ISBN: 978-3-540-69115-0
eBook Packages: Springer Book Archive