Svetlana S. Methods and techniques for circular measurement characteristics accuracy estimation

Thesis is devoted to the directional statistics and their uncertainty estimation accuracy increasing methods and techniques development. Based on symmetrical distributions analysis the expediency of asymmetrical measurement results confidence interval representation is justified. The method for the directional statistics and measurement result uncertainty estimation is developed. The method is based on random circles empirical distribution Johnson curves approximation. Statistical estimates and the confidence interval limits are obtained by the fitted probability density function analytical form. The technique based on bootstrap methods for the main directional statistics such as mean, median, mode and their expanded uncertainty is offered. The software based on offered methods and techniques for random circular data modelling and proceeding is developed. The software is supposed to be applied to the random circles computer-measuring experiments and metrological support organization.