The Ninth International Conference "Computer Data Analysis and Modeling: Complex Stochastic Data and Systems" (CDAM'2010) organized in Minsk by the Belarusian State University and Vienna University of Technology on September 7-11, was devoted to the topical problems in computer data analysis and modeling. There were 102 presentations by more than 150 participants from 21 countries.

Simple conditions for the inconsistency of classical goodness-of-fit tests in case of very sparse categorical data are given. The conditions have the following interesting feature of "reversed consistency": the greater deviation from the null hypothesis the less power of the test. In the paper χ 2 -type criterion based on profile statistics is introduced as an alternative to classical tests in the case of sparse categorical data.

For (very) sparse nominal data, common goodness-of-fit tests usually fail. Alternative goodness-of-fit tests based on extended empirical Bayes approach and grouping are proposed and their consistency is proved. The performance of the tests is illustrated and compared with classical criteria by Monte Carlo simulations.

In this paper, we address the problem of an efficient estimation strategy when solving real survey sampling problems. We consider the panel data model with random effects as a superpopulation model. We investigate the performance of model-assisted estimators under model-based sample designs in a simulation experiment based on the real data (Lithuanian survey on short-term statistics on service) taken as a true population. Here the model-based sampling refers to a sample design constructed using the model and information about the behavior of the model-assisted estimator on available auxiliary data.

In this paper, we derive two strongly consistent Hurst index estimators from the limit behavior of the first- and second-order quadratic variations of the solution of an SDE driven by a fractional Brownian motion. We also show that these estimators retain their properties if the solution is replaced with its Milstein approximation.