| |||||||
Additional SoftwareSuite of S-Plus Functions Recurrent.sThe function recurrent.s implements the semi-parametric analysis of rate and mean functions for recurrent events described by Lawless and Nadeau (Techometrics, 1995, 37, pp.158-168); see also Lin et al. (J. Roy. Statist., Soc. B, 2000, 62, pp.711-730) and Nelson (Technometrics, 1995, 37, pp.147-157). These methods are described in Sections 3.4 and 3.6 of R.J. Cook and J.F. Lawless, "The Statistical Analysis of Recurrent Events" (Springer, 2007) to which equation and solution numbers below refer. In particular, recurrent.s can calculate (i) a Nelson-Aalen estimate of the mean function (3.17) along with Poisson-based variance estimates (3.18) and robust variance estimates (3.34), (ii) the estimate of regression coefficient beta given in Section 3.4.2, the corresponding estimate (3.24) of the baseline mean function Mo(t), and both Poisson-based and robust variance estimates, described in Sections 3.4.2 and 3.6.3, respectively. The functionality of recurrent.s with large data sets depends on the operating system, physical and virtual memory, and the available hard drive space of the machine. An option involving piecewise-constant rate functions is also provided, which can be used even with very large data sets. In this case cut-points for the rate function must be specified; see the NOTES ON INPUT VARIABLES below, and Section 3.3 of Cook and Lawless (2007). Program File Notes for Users | |||||||
|
The R Project for Statistical Computing
|