lm_eps: Linear regression for extreme phenotype sampling data under...

In xu1912/EPSLASSO: Sparse Regression for data under extreme phenotype sampling

Linear regression for extreme phenotype sampling data under low-dimensional situation (sample size n > p number of predictors).

Description

This function solves the Maximum Likelihood Estimate of the low-dimensional linear model for extreme phenotype sampling data using Newton-Raphson (NR) procedure. This function is prepared based on functions from the R package CEPSKAT.

Usage

lm_eps(formula, c1, c2, delta = 0.001, MAXITERNUM = 1000, data = NULL,
  verbose = FALSE)

Arguments

formula	Regression model to be fit. Required.
c1	Right censored point. Required.
c2	Left censored point. Required.
delta	Convergence threshold for NR procedure. Default is 0.001.
MAXITERNUM	Maximum iteration number for NR procedure. Default is 1000.
data	The dataframe stores data for the formula. Default is NULL.
verbose	Print debugging info or not. Default is FALSE.

Examples

n=100
p1=0.2
p2=0.2
X=rnorm(n)
Y=1+0.5*X+rnorm(n)

Y_eps=Y[order(Y)[c(1:(n*p1),(n-n*p2+1):n)]]
X_eps=X[order(Y)[c(1:(n*p1),(n-n*p2+1):n)]]
c1=Y_eps[n*p1+1]
c2=Y_eps[n*p1]

res=lm_eps(Y_eps~X_eps, c1, c2)
res

xu1912/EPSLASSO documentation built on May 10, 2022, 11:23 a.m.