Local Regression and Likelihood, Figure 7.1. Estimating the hazard rate for censored survival (or failure time) data. Hazard rate estimation is specified by 'family','hazard'. The censoring indicator variable is passed as the 'cens' argument. This should be a vector of 0's and 1's, with 1 indicating a censored observation, 0 uncensored. The 'xlim' argument specifies bounds on the domain of the x variable (survival times). Usually, survival times are non-negative (lower bound 0), but with theoretically no upper bound. The specification [0;10000] gives the lower bound of 0, and the upper bound is effectively infinite. Author: Catherine Loader NEEDS: m argument to lfplot. Also, lfplot symbols should distinguish between censored and uncensored data points.

0001 % Local Regression and Likelihood, Figure 7.1. 0002 % 0003 % Estimating the hazard rate for censored survival (or 0004 % failure time) data. Hazard rate estimation is specified 0005 % by 'family','hazard'. 0006 % 0007 % The censoring indicator variable is passed as the 'cens' 0008 % argument. This should be a vector of 0's and 1's, with 1 0009 % indicating a censored observation, 0 uncensored. 0010 % 0011 % The 'xlim' argument specifies bounds on the domain of the 0012 % x variable (survival times). Usually, survival times are 0013 % non-negative (lower bound 0), but with theoretically no 0014 % upper bound. The specification [0;10000] gives the lower 0015 % bound of 0, and the upper bound is effectively infinite. 0016 % 0017 % Author: Catherine Loader 0018 % 0019 % NEEDS: m argument to lfplot. 0020 % Also, lfplot symbols should distinguish between censored and 0021 % uncensored data points. 0022 0023 load heart; 0024 fit = locfit(surv,'cens',cens,'family','hazard','alpha',0.4,'xlim',[0;10000]); 0025 figure('Name','fig7_1: Hazard rate for censored data' ); 0026 lfplot(fit); 0027 xlabel('Survival Time'); 0028 ylabel('Hazard Rate');

Generated on Fri 28-Sep-2012 12:34:30 by