Sign in
Influence diagnostics on competing risks using Cox'smodel with censored data
Journal article

Influence diagnostics on competing risks using Cox'smodel with censored data

F Elfaki, Bin Daud I, N Ibrahim, M Abdullah, M Usman and Mohamed G. Elfaki
WSEAS Transactions Mathematics, Vol.5(1), pp.12-17
01/01/2006

Abstract

This paper studies influence diagnostics (Cook's Distance and Likelihood Distance) on competing risk models when true covariates are observable. The one-step EM and one-step ML methods are used to detect influential observations for Cox's model. It was found that results generated from data analysis using one-step EM algorithm are better than results obtained by using one-step ML method. Moreover, Cook's distance show better results compare to likelihood distance.

Metrics

1 Record Views

Details