Spline Estimation of Discontinuous Regression Functions

Abstract This article deals with regression function estimation when the regression function is smooth at all but a finite number of points. An important question is: How can one produce discontinuous output without knowledge of the location of discontinuity points? Unlike most commonly used smoothers that tend to blur discontinuity in the data, we need to find a smoother that can detect such discontinuity. In this article, linear splines are used to estimate discontinuous regression functions. A procedure of knot-merging is introduced for the estimation of regression functions near discontinuous points. The basic idea is to use multiple knots for spline estimates. We use an automatic procedure involving the least squares method, stepwise knot addition, stepwise basis deletion, knot-merging, and the Bayes information criterion to select the final model. The proposed method can produce discontinuous outputs. Numerical examples using both simulated and real data are given to illustrate the performance of th...

[1]  K. Gordon,et al.  Modeling and Monitoring Biomedical Time Series , 1990 .

[2]  C. J. Stone,et al.  Additive Regression and Other Nonparametric Models , 1985 .

[3]  F. Lombard.  Detecting change points by Fourier analysis , 1988 .

[4]  P. Hall,et al.  Edge-preserving and peak-preserving smoothing , 1992 .

[5]  W. Press,et al.  Numerical Recipes in Fortran: The Art of Scientific Computing.@@@Numerical Recipes in C: The Art of Scientific Computing. , 1994 .

[6]  Heping Zhang,et al.  Maximal Correlation and Adaptive Splines , 1994 .

[7]  Carl de Boor,et al.  A Practical Guide to Splines , 1978, Applied Mathematical Sciences.

[8]  D. Hinkley.  Inference about the intersection in two-phase regression , 1969 .

[9]  J. Friedman.  Multivariate adaptive regression splines , 1990 .

[10]  D. Donoho.  On Minimum Entropy Segmentation , 1994 .

[11]  G. Cobb.  The problem of the Nile: Conditional solution to a changepoint problem , 1978 .

[12]  Yazhen Wang.  Jump and sharp cusp detection by wavelets , 1995 .

[13]  Steve C. A. Peters,et al.  Comparing Automatic Smoothers (A Public Service Enterprise) , 1992 .

[14]  Leo Breiman,et al.  Fitting additive models to regression data , 1993, Computational Statistics & Data Analysis.

[15]  George A. F. Seber,et al.  Linear regression analysis , 1977 .

[16]  I. Johnstone,et al.  Ideal spatial adaptation by wavelet shrinkage , 1994 .

[17]  H. Müller.  CHANGE-POINTS IN NONPARAMETRIC REGRESSION ANALYSIS' , 1992 .

[18]  Naomi Altman,et al.  Kernel Smoothing of Data with Correlated Errors , 1990 .

[19]  Edward Carlstein,et al.  Change-point problems , 1994 .

[20]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[21]  Bernard W. Silverman,et al.  The discrete wavelet transform in S , 1994 .

[22]  J. Friedman,et al.  FLEXIBLE PARSIMONIOUS SMOOTHING AND ADDITIVE MODELING , 1989 .

[23]  John Alan McDonald,et al.  Smoothing with split linear fits , 1986 .

[24]  Y. Yin,et al.  Detection of the number, locations and magnitudes of jumps , 1988 .

[25]  D. Hinkley.  Inference in Two-Phase Regression , 1971 .

[26]  Stéphane Mallat,et al.  Singularity detection and processing with wavelets , 1992, IEEE Trans. Inf. Theory.

[27]  Youngjo Lee,et al.  Bivariate B-splines in generalised linear models , 1994 .