UMRhess

UMRhess_generic

Current (unsorted) estimate/iterate at which to compute
gradient. (Length is &lt;= than the number of X observations in the
problem).

Weights (nonnegative, sum to 1) corresponding to mm. Same
length as mm.

ww_m

Y (response) observation vector (numeric)

Weights (nonnegative, sum to 1) corresponding to yy. Same
length as yy. Default is just 1/length(yy) for each value.

ww_y

This is the error density, a function object (Balabdaoui, Doss, Durot (2020+). Function accepting vector or matrix arguments.

dens

This is derivative of "B" function ("B prime"), where B is
defined in the paper. Function accepting vector or matrix arguments.

Compute Hessian of Unlinked Monotone Regression objective function
 from Balabdaoui, Doss, and Durot

Unmatched regression refers to the regression setting where
covariates and predictors are collected separately/independently and so are not paired together, as in the usual regression setting. Balabdaoui, Doss, and Durot (2021) <arXiv:2007.00830> study the unmatched regression setting where the univariate regression function is known to be monotone. This package implements methods for computing the estimator developed in Balabdaoui, Doss, and Durot (2021). The main method is an active-set-trust-region-based method.

UMRhess: Compute Hessian of Unlinked Monotone Regression objective function from Balabdaoui, Doss, and Durot

Description

Usage

Arguments

Details