Dr. Russell Stocker of the Mathematics Department and Dr. Akim Adekpedjou had their manuscript “Optimal Goodness-of-Fit Tests for Recurrent Event Data” recently accepted for publication in the journal Lifetime Data Analysis.
The article proposes a class of goodness-of-fit tests for assessing the parametric form of the baseline intensity process found in models used to analyze recurrent event data. These tests are based on a class of weighted empirical processes. Asymptotic properties of this class of weighted empirical processes are given for a sequence of Pitman alternatives. Test statistics are constructed as functionals of the weighted empirical processes. Optimal choices for the weight process are found for a class of chi-squared tests. Khmaladze's transformation is applied to the test statistic, and the asymptotic null distribution of the resulting transformation is given. Two data sets are used to illustrate the goodness-of-fit tests.