iteration.fun.cox: One-step iteration function for divide-and-conquer Cox model

In michaelyanwang/divideconquer: Divide-and-conquer adaptive lasso using least square approximation

Description

iteration.fun.cox estimates a one-step for a Cox regression for each subset.

Usage

iteration.fun.cox(dat.list, bini, kk.list, rtn = "Score+A+Approx")

Arguments

dat.list	list of subsets (after dividing). In each subset, first column = U, second column = delta (0/1), rest = design matrix
bini	initial estimator as starting point
kk.list	which subsets of dat.list get one-step update
rtn	return score, score+negative information, score+negative information using -SS' approximation

Value

a list with b.k a matrix of one-step estimators and with Ahat the negative information matrix

Author(s)

Yan Wang, Tianxi Cai

Examples

iteration.fun.cox(dat.list=dat.list,bini=bini,kk.list=2:K,rtn='Score+A+Approx')

michaelyanwang/divideconquer documentation built on Aug. 16, 2019, 10:11 a.m.