Hostname: page-component-745bb68f8f-f46jp Total loading time: 0 Render date: 2025-01-07T18:56:52.019Z Has data issue: false hasContentIssue false
  • English
  • Français

Monotone Regression: Continuity and Differentiability Properties

Published online by Cambridge University Press:  01 January 2025

J. B. Kruskal
J. B. Kruskal*
Affiliation:
Bell Telephone Laboratories, Incorporated, Murray Hill, New Jersey
Get access Rights & Permissions [Opens in a new window]

Abstract

Least-squares monotone regression has received considerable discussion and use. Consider the residual sum of squares Q obtained from the least-squares monotone regression of yi on xi. Treating Q as a function of the yi, we prove that the gradient ▽Q exists and is continuous everywhere, and is given by a simple formula. (We also discuss the gradient of d = Q1/2.) These facts, which can be questioned (Louis Guttman, private communication), are important for the iterative numerical solution of models, such as some kinds of multidimensional scaling, in which monotone regression occurs as a subsidiary element, so that the yi and hence indirectly Q are functions of other variables.

Keywords

Public PolicyStatistical TheoryPrivate CommunicationMultidimensional ScalingSimple Formula
Type
Original Paper
Information
Psychometrika , Volume 36 , Issue 1 , March 1971 , pp. 57 - 62
Copyright
Copyright © 1971 The Psychometric Society

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

Ayer, M., Brunk, H. D., Ewing, G. M., Reid, W. T., & Silverman, Edward An empirical distribution function for sampling with incomplete information. Annals of Mathematical Statistics, 1955, 26, 641–647CrossRefGoogle Scholar
Bartholomew, D. J. A test of homogeneity for ordered alternatives. Biometrika, 1959, 46, 36–48CrossRefGoogle Scholar
Bartholomew, D. J. A test of homogeneity of means under restricted alternatives (with discussion). Journal of the Royal Statistics Society, Series B, 1961, 23, 239–281CrossRefGoogle Scholar
Barton, D. E., & Mallows, C. L. The randomization bases of the amalgamation of weighted means. Journal of the Royal Statistics Society, Series B, 1961, 23, 423–433CrossRefGoogle Scholar
Kruskal, J. B. Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika, 1964, 29, 1–28CrossRefGoogle Scholar
Kruskal, J. B. Nonmetric multidimensional scaling: a numerical method. Psychometrika, 1964, 29, 115–129CrossRefGoogle Scholar
Kruskal, J. B. Analysis of factorial experiments by estimating monotone transformations of the data. Journal of the Royal Statistical Society, Series B, 1965, 27, 251–263CrossRefGoogle Scholar
Kruskal, J. B. Two convex counterexamples: a discontinuous envelope function and a non-differentiable nearest-point mapping, accepted by Proceedings of the American Math. Society, 1969a.CrossRefGoogle Scholar
Kruskal, J. B. & Carmone, Frank MONANOVA, a Fortran IV program for monotone analysis of variance. Journal of Marketing Research (to appear).Google Scholar
Kruskal, J. B., & Carmone, Frank MONANOVA: A Fortran IV program for monotone analysis of variance, (non-metric analysis of Factorial Experiments). Behavioral Science, 1969, 14, 165–166Google Scholar
Kruskal, J. B., & Carroll, J. D. Geometrical models and badness-of-fit functions. In Krishnaiah, P. R. (Eds.), Multivariate Analysis—II, 1969, New York City: Academic PressGoogle Scholar
Kruskal, J. B. A new convergence condition for methods of ascent, submitted to the Short Notes section of SIAM Review 1970.Google Scholar
Miles, R. E. The complete amalgamation into blocks, by weighted means, of a finite set of real numbers. Biometrika, 1959, 46, 317–327CrossRefGoogle Scholar
Moreau, Jean Jaques Convexity and duality. In Caianiello, E. R. (Eds.), Functional analysis and optimization, 1969, New York: Academic PressGoogle Scholar
Nashed, M. Z. A decomposition relative to convex sets. Proceedings of the American Mathematical Society, 1968, 19, 782–786CrossRefGoogle Scholar
van Eeden, C. Maximum likelihood estimation of partially or completely ordered parameters, I. Proceedings of Akademie van Wetenschappen, Series A, 1957, 60, 128–136Google Scholar
van Eeden, C. Note on two methods for estimating ordered parameters of probability distributions. Proceedings of Akademie van Uetenschappen, Series A, 1957, 60, 506–512Google Scholar
Witsenhausen, H. A minimax control problem for sampled linear systems. IEEE Transactions on Automatic Control, 1968, AC-13, 5–21CrossRefGoogle Scholar
Young, F. W., & Torgerson, W. S. TORSCA, a Fortran IV program for Shepard-Kruskal multidimensional scaling analysis. Behavioral Science, 1967, 12, 498–498Google Scholar