B1511
Title: IndTestPP: An R package for testing independence between point processes in time
Authors: Ana C Cebrian - University of Zaragoza (Spain) [presenting]
Jesus Abaurrea - University of Zaragoza (Spain)
Jesus Asin - University of Zaragoza (Spain)
Abstract: IndTestPP (Tests of Independence Between Point Processes in Time) is an R package aiming to provide different parametric and nonparametric tests to check the independence between two or more homogeneous or nonhomogeneous point processes. The following tools are implemented: two tests for Poisson processes, a conditional test and a parametric bootstrap test based on the close point set. The last one is also applicable to any parametric point process which can be simulated. A nonparametric Lotwick-Silverman test, also based on the close point set, is developed and it is applicable to any homogeneous point process, without any distribution model assumption. Versions of the spatial $K$ and $J$-function adapted to time processes are developed, together with nonparametric Lotwick-Silverman tests based on those measures. A graphical procedure, the Dutilleul plot for two point processes, is also implemented. The package provides tools for generating trajectories of a vector of time point processes with different types of dependence: common Poisson shock processes, networks of queues, multivariate Neyman-Scott process with dependent cluster centers, and marked Poisson process with dependent marks generated by a Markov chain. Different examples of use will be shown.
Journal