= Breslow approximation for tied event times
{c}
{title2=$\frac{\exp(\sum_{i\in D_j}x_i^T\beta)}{(\sum_{i\in R_j}e^{x_i^T\beta})^{d_j}}$}
When $d_j$ events occur at the same recorded time, the Breslow approximation to <Cox partial likelihood> uses the factor $\exp(\sum_{i\in D_j}x_i^T\beta)/(\sum_{i\in R_j}e^{x_i^T\beta})^{d_j}$, where $R_j$ is the <risk set> and $D_j$ the event set. It approximates the ordering of tied events rather than treating their within-tie order as known.
Back to article page
