B1839
Title: Sample cross-covariance function for testing of bi-dimensional Gaussian processes
Authors: Katarzyna Maraj-Zygmat - Wroclaw University of Science and Technology (Poland) [presenting]
Abstract: A novel framework is introduced that allows efficient bi-dimensional Gaussian process discrimination. The underlying test statistic is based on the sample cross-covariance function. There is a lack of methods for these processes. We present the analysis of probabilistic properties of the sample cross-covariance function for selected multivariate Gaussian processes (bi-dimensional Brownian motion, bi-dimensional fractional Brownian motion, and bi-dimensional Ornstein-Uhlenbeck process). Also, the test based on the sample cross-covariance function for multivariate Gaussian processes is presented. The results and properties are checked by Monte Carlo simulation. For completeness, it is also shown how to embed this methodology into bi-dimensional real data analysis.
