看板 Statistics 關於我們 聯絡資訊
題目上不太懂怎麼做 因為不懂s的分佈 為甚麼會N(σ,σ^2/2(n-1)) Normal的話 那是用Z嗎 還是用t@@ It is known thae the point estimator toward population standard deviation σ is the sample standard deviation s.This estimator is biased and has a complicated sampling distribution.However, when the sample size n is large, the sampling distribution of s can be shown to follow N(σ,σ^2/2(n-1)) approximately. (1) Construct a formula of the 100(1-α)% confidence interval for σ in this case. (2)Sometimes, it is desired to control the sampling error within a level e by choosing an appropriate sample size.With a planning value for σ from past experience and a confidence coefficient 1-α,derive the minimum sample size required. (3)Consider the hypothesis test about σ. If we feel that current σ would be smaller from past experience,what are the appropriate hypotheses? With type I risk specified as α,what is the critical region to reject null hypothesis(in the form of s>c or s<c)? -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 220.133.229.161 ※ 編輯: selient 來自: 220.133.229.161 (08/30 17:42)
cyhuangjhu:題目出得不錯阿.....想想如果題目中的s改成平均值 08/30 22:14
cyhuangjhu:那就照著解這題 只是asymptotic distribution 的 08/30 22:21
cyhuangjhu:variance 的estimate 跟s有關 08/30 22:22