Browse files

Add discussion of Nelson-Aalen

  • Loading branch information...
1 parent 0c22b9c commit 4767e7e139108b216b6b5af1e7bd4576c179ac6a @patperry committed May 30, 2012
Showing with 6 additions and 0 deletions.
  1. +6 −0 iproc.tex
@@ -1602,6 +1602,12 @@ \section{Conclusion}\label{S:discussion}
continuous-time information we get precise quantification on the
time-dependent behavior of the network effects.
+In this work, the focus has been on the coefficient vector $\beta$. Since we
+have used partial likelihood for the estimation, it has not been necessary to
+the baseline intensity $\bar \lambda_t(i)$. Were we to use the model for
+prediction, we would need to estimate the baseline intensities. This could be
+done using a Nelson-Aalen estimator as in~\citet{andersen1993statistical}.
The foundation of our work is Cox's \citeyearpar{cox1972regression}
proportional intensity model and partial likelihood theory, tools
which he first introduced almost forty years ago and which have been

0 comments on commit 4767e7e

Please sign in to comment.