On the fitting of the three-parameter distributions Lognormal, Gamma, and Weibull

摘要: This paper deals with the problem of fitting a theoretical, continuous, univariate distribution to an empirical one. Analyzing such distributions we often assume Gauss distribution. is not satisfying in case asymmetric which show lower bound, for instance left-asymmetry cycletime work-study. For three-parameter families Lognormal, Gamma, and Weibull are suitable. In first step, each family estimate three parameters out sample data by means maximum likelihood method (scoring method). second step discriminate between ratio tests. The unknown test statistics simulated computer. whole procedure included application programme illustrated example. statistical instrument analyzing bound as they exist work-study, hydrology (maximum drainage water flow), many other fields.