- 최우 추정기 - 랜덤 샘플의 함수이기 때문에 (고정되지 않은)도 랜덤이다. 의 표준 오차의 추정 α는 , 피셔 정보에서 얻을 수있다α^α^
I(θ)=−E[∂2L(θ|Y=y)∂θ2|θ]
Where θ is a parameter and L(θ|Y=y) is the log-likelihood function of θ conditional on random sample y.
Intuitively, the Fisher information indicates the steepness of the curvature of the log-likelihood surface around the MLE, and so the amount of 'information' that y provides about θ.
For a Pareto(α,y0) distribution with a single realization Y=y, the log-likelihood where y0 is known:
L(α|y,y0)L′(α|y,y0)L′′(α|y,y0)=logα+αlogy0−(α+1)logy=1α+logy0−logy=−1α2
Plugging in to the definition of Fisher information,
I(α)=1α2
For a sample
{y1,y2,...,yn} The maximum likelihood estimator
α^ is asymptotically distributed as:
α^∼n→∞N(α,1nI(α))=N(α,α2n),
Where
n is the sample size. Because
α is unknown, we can plug in
α^ to obtain an estimate the standard error:
SE(α^)≈α^2/n−−−−√≈4.69312/5−−−−−−−−√≈2.1