Methods of the sequential testing of hypotheses in problems of analysis of random processes involving structural changes

摘要

Algorithms of detecting changes in the properties of random processes are considered, which are intended for the investigation and monitoring of economic processes. Possibilities of the use of the hyperbolic distribution for the description of these processes are examined. Algorithms are developed and investigated of the detection of a disorder for hyperbolic distributions in the case of exactly known parameters before and after the disorder and in the case where a domain in the space of parameters is specified after the disorder. A comparison of the normal and the hyperbolic distribution is made for the description and detection of changes in financial indices.