威布尔分布尺度参数的最佳双边检验

众所周知,假设检验中有两类错误,即弃真和受伪的错误.传统检验方法拒绝域临界点的选取都是以尾部概率相等为准,这种方法得到的拒绝域使得犯第二类错误的概率累积值并非最小.此处从理论上推求给出威布尔分布尺度参数在犯第二类错误的概率累积最小意义下的最佳双边检验,并结合实例说明了最佳双边检验较传统双边检验的优越性.

Optimal Two-side Test of Weibull Distribution Scale Parameter

It is well-known that there are two classes of errors in hypothesis test, i.e., disposable true and pseudo error.The rejection of region critical point selection in traditional test methods is all based on tail probability equality, and the rejection region obtained by this method makes the probability accumulation value for making the second class of error not minimum. The optimal two-side test of Weibull distribution scale parameter under the minimum sense of the probability accumulation for making the second class of error is given by theoretical deduction and solution. Examples are used to explain the advantage of the optimal two-side test over traditional two-side test.