lrm.fit.bare: lrm.fit.bare

In rms: Regression Modeling Strategies

lrm.fit.bare

Description

Bare Bones Logistic Regression Fit

Usage

lrm.fit.bare(x, y, maxit = 12, eps = 0.025, tol = 1e-07)

Arguments

x	a vector of matrix of covariate values
y	a numeric or factor vector representing the dependent variable
maxit	maximum number of iteractions
eps	stopping criterion (change in -2 log likelihood)
tol	matrix inversion tolerance for singularities

Details

This is a stripped down version of the lrm.fit() function that computes only the regression coefficients, variance-covariance-matrix, and log likelihood (for null and fitted model) and does not compute any model fit indexes etc. This is for speed in simulations or with bootstrapping. Missing data are not allowed. The function handles binary and ordinal logistic regression (proportional odds model).

Value

a list with elements coefficients, var, fail, freq, deviance

Author(s)

Frank Harrell

rms documentation built on Sept. 12, 2023, 9:07 a.m.