一种带判别函数的单样本有限状态分布检验

A One-Sample Finite State Distribution Test with Discriminant Function

本文引进了一种带判别函数的单样本有限状态分布检验,并给出了基于精确分布的检验计算方法。在本文方法中,给出不同的判别函数,检验可表示不同的具体检验方法,其中几个经典的检验方法,例如X2拟合优度检验、二项检验和似然比检验都是本文检验方法在不同判别函数下的特例。本文给出了一个由样本的分布频率和总体的概率的差距构成的判别函数。案例分析显示,这个检验方法比上述三种经典的检验方法都更有效率。

This paper introduces a one-sample finite state distribution test with a discriminant function, and gives a test calculation method based on exact distribution. In the method of this paper, given different discriminant functions, the test can represent different specific test methods, among which several classic test methods, such as Х2 goodness-of-fit test, binomial test, and likelihood ratio test are all tested in this paper. This paper presents a discriminant function composed of the difference between the distribution frequency of the sample and the probability of the population. The case study shows that this test method is more efficient than the three classical test methods mentioned above.