安徽大学学报(自然科学版)
安徽大學學報(自然科學版)
안휘대학학보(자연과학판)
JOURNAL OF ANHUI UNIVERSITY(NATURAL SCIENCES EDITION)
2010年
6期
7-12
,共6页
有限混合总体%广义似然比%检验统计量%渐近分布%检验功效
有限混閤總體%廣義似然比%檢驗統計量%漸近分佈%檢驗功效
유한혼합총체%엄의사연비%검험통계량%점근분포%검험공효
finite mixture population%GLR%test statistics%asymptotical distribution%test power
在正则统计模型中,广义似然比检验其方法简洁而有效.但在经济、生物学等领域中有着广泛的应用的混合总体中,因其正则条件不再成立,使得广义似然比检验要困难复杂得多.概述了在混合模型中广义似然比检验研究的进展,特别是检验统计量渐近分布的新进展.
在正則統計模型中,廣義似然比檢驗其方法簡潔而有效.但在經濟、生物學等領域中有著廣汎的應用的混閤總體中,因其正則條件不再成立,使得廣義似然比檢驗要睏難複雜得多.概述瞭在混閤模型中廣義似然比檢驗研究的進展,特彆是檢驗統計量漸近分佈的新進展.
재정칙통계모형중,엄의사연비검험기방법간길이유효.단재경제、생물학등영역중유착엄범적응용적혼합총체중,인기정칙조건불재성립,사득엄의사연비검험요곤난복잡득다.개술료재혼합모형중엄의사연비검험연구적진전,특별시검험통계량점근분포적신진전.
In the regular statistical model, GLR (generalized likelihood ratio) tests were simple and had optimal test power in some sense. But in the finite mixture models, which were applicable in economics and biology, GLR tests were very complex and difficult because some regular conditions were not satisfied. In this paper, we reviewed some development of generalized likelihood ratio test, especially about asymptotical distribution of GLR test statistics under null hypothesis in mixture models.