CMLS-package {CMLS} | R Documentation |
Constrained Multivariate Least Squares
Description
Solves multivariate least squares (MLS) problems subject to constraints on the coefficients, e.g., non-negativity, orthogonality, equality, inequality, monotonicity, unimodality, smoothness, etc. Includes flexible functions for solving MLS problems subject to user-specified equality and/or inequality constraints, as well as a wrapper function that implements 24 common constraint options. Also does k-fold or generalized cross-validation to tune constraint options for MLS problems. See ten Berge (1993, ISBN:9789066950832) for an overview of MLS problems, and see Goldfarb and Idnani (1983) <doi:10.1007/BF02591962> for a discussion of the underlying quadratic programming algorithm.
Details
The DESCRIPTION file:
Package: | CMLS |
Type: | Package |
Title: | Constrained Multivariate Least Squares |
Version: | 1.0-1 |
Date: | 2023-03-29 |
Author: | Nathaniel E. Helwig <helwig@umn.edu> |
Maintainer: | Nathaniel E. Helwig <helwig@umn.edu> |
Depends: | quadprog, parallel |
Description: | Solves multivariate least squares (MLS) problems subject to constraints on the coefficients, e.g., non-negativity, orthogonality, equality, inequality, monotonicity, unimodality, smoothness, etc. Includes flexible functions for solving MLS problems subject to user-specified equality and/or inequality constraints, as well as a wrapper function that implements 24 common constraint options. Also does k-fold or generalized cross-validation to tune constraint options for MLS problems. See ten Berge (1993, ISBN:9789066950832) for an overview of MLS problems, and see Goldfarb and Idnani (1983) <doi:10.1007/BF02591962> for a discussion of the underlying quadratic programming algorithm. |
License: | GPL (>=2) |
Index of help topics:
CMLS-package Constrained Multivariate Least Squares IsplineBasis I-Spline Basis for Monotonic Polynomial Splines MsplineBasis M-Spline Basis for Polynomial Splines cmls Solve a Constrained Multivariate Least Squares Problem const Print or Return Constraint Options for cmls cv.cmls Cross-Validation for cmls mlsei Multivariate Least Squares with Equality/Inequality Constraints mlsun Multivariate Least Squares with Unimodality (and E/I) Constraints
The cmls
function provides a user-friendly interface for solving the MLS problem with 24 common constraint options (the const
function prints or returns the different contraint options). The cv.cmls
function does k-fold or generalized cross-validation to tune the constraint options of the cmls
function. The mlsei
function solves the MLS problem subject to user-specified equality and/or inequality (E/I) constraints on the coefficients. The mlsun
function solves the MLS problem subject to unimodality constraints and user-specified E/I constraints on the coefficients.
Author(s)
Nathaniel E. Helwig <helwig@umn.edu>
Maintainer: Nathaniel E. Helwig <helwig@umn.edu>
References
Goldfarb, D., & Idnani, A. (1983). A numerically stable dual method for solving strictly convex quadratic programs. Mathematical Programming, 27, 1-33. doi:10.1007/BF02591962
Helwig, N. E. (in prep). Constrained multivariate least squares in R.
Ten Berge, J. M. F. (1993). Least Squares Optimization in Multivariate Analysis. Volume 25 of M & T Series. DSWO Press, Leiden University. ISBN: 9789066950832
Turlach, B. A., & Weingessel, A. (2019). quadprog: Functions to solve Quadratic Programming Problems. R package version 1.5-8. https://CRAN.R-project.org/package=quadprog
Examples
# See examples for cmls, cv.cmls, mlsei, and mlsun