Testing the Reliability of Parallel Systems Using Generalized p Values
The concept of a generalized p-value is applied for testing one-sided hypotheses of reliability of parallel systems. The generalized test variable is obtained by the method of fiducial inference, and we have a conclusion that the generalized p-value is the fiducial probability of the null hypothesis. The true size of the test with fixed level of significance is investigated. It is shown that the test is exact, i.e., it provides a size-guaranteed inference. Simulation results show that the performance of the actual size and power is quite satisfactory.
Reliability of Parallel Systems Ezponential Distribution Generalized p-value Fiducial Inference
MA Jianjun XU Xingzhong CONG Peijuan
School of Science, Shenyang Institute of Technology, P.R.China, 110168 Department of Mathematics, Beijing Institute of Technology, P.R.China, 100081
国际会议
2009 International Institute of Applied Statistics Studies(2009 国际应用统计学术研讨会)
青岛
英文
1-6
2009-07-25（万方平台首次上网日期，不代表论文的发表时间）