Analysis and Experiment of Two Kind Errors in Hypothesis Test
Yali Lu · Journal of Henan Institute of Education · 2011
Based on minor probability principle,central-limit theorem,and sampling-distribution theory,studied relations depending on quantity variation among the probability α of violating-TRUE mistakes,the probability β of taking-FALSE mistakes,and sample size n in hypothesis test.Specific expression among them was also given.In addition,designed numerical experiment of significant test and gave the corresponding MATLAB code.Results show that for the fixed n,the α's loss is β's gain.For the fixed α and in order to decrease β,one must increase sample size n.By estimating absolute error between the expectation value and actual value of ensemble,the probability of taking-FALSE mistakes can be approximately calculated.The frequency of taking-FALSE mistakes obtained by numerical experiments is in accord with the result calculated via mathematical formula.