jac_est_fun: Jacobian of PIM U-estimating function

In ge-li/upim: U-statistics Probabilistic Index Models

Jacobian of PIM U-estimating function

Description

The Jacobian has the generic form \dot{δ}(b) = \frac{\partial}{\partial b^{T}} δ(b) = \frac{≤ft(2 μ y-y-μ^{2}\right) \dot{μ}^{2}+μ(1-μ)(y-μ) \ddot{μ}}{μ^{2}(1-μ)^{2}} x x^{T}.

Usage

jac_est_fun(y, X, b, link, w = NULL)

Arguments

y	numeric A vector of pseudo observations.
X	matrix The pairwise difference of original design matrix.
b	numeric The PIM parameter.
link	character The link function: "logit", or "probit".
w	numeric The weights, default is NULL.

Value

matrix A p by p Jacobian matrix of res_fun(...).

ge-li/upim documentation built on Oct. 30, 2022, 3:36 a.m.