作者: M. A. Stephens
DOI: 10.1080/01621459.1974.10480196
关键词:
摘要: Abstract This article offers a practical guide to goodness-of-fit tests using statistics based on the empirical distribution function (EDF). Five of leading are examined—those often labelled D, W 2, V, U A 2—and three important situations: where hypothesized F(x) is completely specified and represents normal or exponential with one more parameters be estimated from data. EDF easily calculated, require only line significance points for each situation. They also shown competitive in terms power.