Parameters | Predictive Modeling | Perform Buckley-James Estimation

Perform Buckley-James Estimation
Check this box to perform survival analysis using the specified model and assuming some of the observations are right-censored.
Estimation is performed by iteratively fitting the model with an updated dependent variable that replaces censored observations with estimates of their true time-to-event values.
The method is a generalization of the original regression algorithm of Buckley and James (1979)1 as expanded upon by Glasson (2007, RMIT PhD Thesis).
Note: You must specify the time-to-event dependent variable (this variable can be log-transformed) as the Dependent Variable on the General tab. You must also specify the Censor Variable and other options on this tab.
To Perform Cross Validation Model Comparison:
*

1
Buckley, J. and I. James. 1979. Linear regression with censored data. Biometrika 66: 429-436.