There have been numerous tests proposed to determine whether or not the exponential model is suitable for a given data set. In this article, we propose a new test statistic based on spacings to test whether the general progressive Type-II censored samples are from exponential distribution. The null distribution of the test statistic is discussed and it could be approximated by the standard normal distribution. Meanwhile, we propose an approximate method for calculating the expectation and variance of samples under null hypothesis and corresponding power function is also given. Then, a simulation study is conducted. We calculate the approximation of the power based on normality and compare the results with those obtained by Monte Carlo simulation under different alternatives with distinct types of hazard function. Results of simulation study disclose that the power properties of this statistic by using Monte Carlo simulation are better for the alternatives with monotone increasing hazard function, and otherwise, normal approximation simulation results are relatively better. Finally, two illustrative examples are presented.

已经提出了许多测试来确定指数模型是否适用于给定的数据集。在本文中，我们提出了一种新的基于间距的检验统计量来检验一般渐进式 II 型删失样本是否来自指数分布。讨论了检验统计量的零分布，它可以近似为标准正态分布。同时，我们提出了一种在原假设下计算样本期望和方差的近似方法，并给出了相应的幂函数。然后，进行模拟研究。我们基于正态性计算功率的近似值，并将结果与​​蒙特卡罗模拟在具有不同类型的危险函数的不同替代方案下获得的结果进行比较。仿真研究结果表明，使用Monte Carlo仿真的该统计量的幂特性对于具有单调递增危险函数的备选方案更好，反之，正态逼近仿真结果相对较好。最后，给出了两个说明性示例。