Nikitin Ya Yu, Ragozin I A
Department of Mathematics and Mechanics, Saint-Petersburg State University, Saint-Petersburg, Russia.
Department of Mathematics and Business Informatics, National Research University - Higher School of Economics, Saint-Petersburg, Russia.
J Appl Stat. 2020 May 7;47(13-15):2610-2622. doi: 10.1080/02664763.2020.1761952. eCollection 2020.
We construct two -empirical tests for the logistic location family which are based on appropriate characterization of this family using independent exponential shifts. We study the limiting distributions and local Bahadur efficiency of corresponding test statistics under close alternatives. It turns out that the present tests are considerably more efficient than the recently proposed similar tests based on another characterization. The efficiency calculations are accompanied by the simulation of power for new tests together with the previous ones.Both efficiency and power turn out to be very high. Finally we consider the application of our tests to real data example.
我们基于使用独立指数偏移对逻辑斯谛位置族进行的适当刻画，构建了该族的两个实证检验。我们研究了在接近备择假设下相应检验统计量的极限分布和局部巴哈杜尔效率。结果表明，当前的检验比最近基于另一种刻画提出的类似检验效率要高得多。效率计算伴随着新检验与先前检验的功效模拟。结果表明效率和功效都非常高。最后，我们考虑将我们的检验应用于实际数据示例。
