最近在国外教材上看到一题
Consider a sample of Y_1,⋯,Y_50 from a normal distribution N(μ,σ^2).
The 95% confidence interval of μ is given by
Y ̅±t_(n-1) (α/2)S/√n=Y ̅±0.2842×S
Use 5000 simulations from N(1,2) to numerically confirm that the coverage probability of the interval of μ = 1 is about 95%.
Re-do, but with the Ys from a X1^2 distribution. (Hence E(Y) = 1, Var(Y) = 2.). Is the simulated coverage probability close to 95%?
怎样确认置信度是不是95%左右呢？ 我都是用的R程序。
