Summary
A relaxed iterative projection (abb. RIP) algorithm for arbitrary linear equation systems is described. It has favorable properties with respect to many statistical applications. A major advantage is that convergence can be established without restrictions on the system matrix. Hence certain characteristics of the system matrix such as diagonal dominance are not required. As a result RIP fitting can be applied where backfitting tends to fail, e.g. when regression predictors are substantially correlated (problem of multicollinearity respectively concurvity). Convergence under a suitable choice of the relaxation parameter is derived for general n × m system matrices. The RIP solution of typical equation systems is studied with respect to the correct (analytical) solution. Empirical findings for the practical selection of the relaxation parameter are reported.
The financial support of the FWF Austrian Science Fund in the research project “Konvergenz und Numerik des Backfitting-Algorithmus” is greatly acknowledged.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Bibliography
Bjorck, A. (1996) Numerical Methods for Least Squares Problems. SIAM. Philadelphia, PA.
Buja, A., Hastie, T. and Tibshirani, R. (1989) Linear smoothers and additive models (with discussion). Annal. Statist. 17, 453–555.
Dominici, F. et al. (2002) On the use of generalized additive models in time-series studies of air pollution and health. Amer. J. Epidemiol., 156, 193–203.
De la Garza, A. (1951) An Iterative Method for Solving Linear Equations. Oak Ridge Gaseous Diffusion Plant, Rep. K-731, Oak Ridge, TN.
Demmel, J. W. (1997) Applied Numerical Linear Algebra. SIAM, Philadelphia, PA.
Gentle, J. E. (1998) Numerical Linear Algebra for Applications in Statistics. Springer, New York.
Green, P. J. and Silverman, B. W. (1994) Nonparametric Regression and Generalized Linear Models. A Roughness Penalty Approach. Chapman & Hall, London.
Golub, G. H. and van Loan, C. F. (1989, 2nd edition) Matrix Computations. The Johns Hopkins University Press, Baltimore, MD.
Hastie, T. J. and Tibshirani, R. J. (1990) Generalized Additive Models. Chapman and Hall, London.
Maess, G. (1988) Projection methods solving rectangular systems of linear equations. J. Comp. Appl. Math. 24, 107–119.
Murty, K. G. (1983) Linear Programming. Wiley, New York.
Opsomer, J. D. and Ruppert, D. (1997) Fitting a bivariate additive model by local polynomial regression. Annal. Statist., 25, 186–211.
Ramsay T.O., Burnett R.T. and Krewski D. (2003) The effect of concurvity in generalized additive models linking mortality to ambient particulate matter. Epidemiology, 14, 18–23.
Schimek, M. G. (1996). An iterative projection algorithm and some simulation results. In Prat, A. (ed.) COMPSTAT’ 96. Proceedings in Computational Statistics. Physica-Verlag, Heidelberg, 453–458.
Schimek, M. G. (2002) RIP-GAMs and classification trees in quantitative MRI. J. Jpn. Soc. Comp. Statist, 15, 123–134.
Schimek, M. G. & Turlach, B. A. (2000) Additive and generalized additive models. In Schimek, M. G. (ed.) Smoothing and Regression. Approaches, Computation and Application. Wiley, New York, 277–327.
Wood, S. (2004) Stable and efficient multiple smoothing parameter estimation for generalized additive models. J. Amer. Statist. Assoc., 99, 637–686.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Physica-Verlag Heidelberg
About this paper
Cite this paper
Schimek, M.G., Stettner, H. (2006). A Relaxed Iterative Projection Algorithm for Rank-Deficient Regression Problems. In: Sperlich, S., Härdle, W., Aydınlı, G. (eds) The Art of Semiparametrics. Contributions to Statistics. Physica-Verlag HD. https://doi.org/10.1007/3-7908-1701-5_6
Download citation
DOI: https://doi.org/10.1007/3-7908-1701-5_6
Publisher Name: Physica-Verlag HD
Print ISBN: 978-3-7908-1700-3
Online ISBN: 978-3-7908-1701-0
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)