fit.ini: Produce pseudo x and y based on marginal proportion odds...

In denisagniel/smrtr:

Produce pseudo x and y based on marginal proportion odds model fits and quadratic

Description

Produce pseudo x and y based on marginal proportion odds model fits and quadratic

Usage

fit.ini(x, y, xname, yname = "y", p)

Arguments

x	Named matrix of predictors
y	n x m matrix of outcomes
xname	String to identify predictors in x
yname	String to identify outcomes in y
p	Number of predictors (doesn't need to be same as number of columns of x) #'

Value

A list with the following elements:

• pseudoY: pseudo-outcome for quadratic expansion of the likelihood

• betahat: coefficients corresponding to the NPMLE of finite-dimensional parameters

• alphahat: coefficients corresponding to the NPMLE of infinite dimensional parameters

• pseudoX: pseudo-predictors for quadratic expansion

• pseudoX.t: adaptive version of pseudo-predictors

• inv.info: inverse Fisher information from marginal models

• inv.info.beta: just the part of the inverse information corresponding to beta

• beta.inds: indices in the coefficient vectors corresponding to beta

• alpha.inds: indices in the coefficient vectors corresponding to the infinite-dimensional parameter

• b0.inds: indices in the coefficient vectors corresponding to nuisance covariates

• coefs: full coefficient vector

denisagniel/smrtr documentation built on Sept. 17, 2022, 10:23 p.m.