Given Probability density Function is-
F(X,θ)=θe−θ
Let l(X,θ) be the natural logrithm of f(x,θ)
l=log(θe−θ)=logθ+loge−θ=logθ−θ
Variance of unbiased estimator of \theta
V(θ^)≥I(θ)1 −(1)
Where, the Fisher information I(θ) is defined by
I(θ)=−E[dθ2d2l(X,θ)]
Where, E= Expected value of samples
I(θ)=−E(−θ21)
I(θ)=θ2E
From eqs.(1) we have
V(θ^)≥I(θ)1
V(θ^)≥Eθ2
Comments
Leave a comment