Dorozhovets Mykhaylo, Warsza Zygmunt Lech
Methods of upgrading the uncertainty of type A evaluation (2) - Elimination of the influence of autocorrelation of observations and choosing the adequate distribution
It is the continuation of a previous paper, in which limitations of uncertainty type A estimation due international guide GUM, are shortly discussed and method of elimination of influence of unknown systematic components: trend and oscillation, are introduced. Below two next proposals of development and upgrading of the method type A are presented, i.e. taking in estimation the influence of autocorrelation of observation resultants and choosing the adequate type of their probability distribution by ?2 criterion, if is different that normal one, are considered. It is illustrated by numerical examples of normal, rectangular and Laplace distribution. Results are compared to parameters calculated by the actual GUM method. Conclusions and background literature are included.