27.5k views
2 votes
Suppose X₁, X₂,....., Xₙ are i.i.d. random variables with finite mean μ and variance σ² ∈(0,[infinity]). We want to estimate σ² by an estimator σ² Consider σ² = 1/n ∑ⁿᵢ₌₁ (Xᵢ −

Xₙ )² , where Xₙ = 1/n ∑ⁿᵢ₌₁ Xᵢ Prove that σ² → σ² Hint: You can use 1/n ∑ⁿᵢ₌₁ (Xᵢ - μ)² → σ² . Xₙ → X and Yₙ → Y iff (Xₙ,, Yₙ ) → (X,Y), where (Xₙ )ₙ∈ₙ and (Yₙ )ₙ∈ₙ denote sequences of random variables.

User Snyf
by
8.0k points

1 Answer

2 votes

Final answer:

To prove that σ² → σ², we show that the estimator σ² converges in probability to the true value σ² as the sample size increases.

Step-by-step explanation:

To prove that σ² → σ², we need to show that the estimator σ² converges in probability to the true value σ² as the sample size increases.

Let's start by rewriting the estimator σ² = 1/n ∑ⁿᵢ₌₁ (Xᵢ − Xₙ )² in terms of the sample mean Xₙ, which is Xₙ = 1/n ∑ⁿᵢ₌₁ Xᵢ. We can substitute Xₙ in the estimator formula to get σ² = 1/n ∑ⁿᵢ₌₁ (Xᵢ − (1/n ∑ⁿᵢ₌₁ Xᵢ))².

Next, we need to use the property (Xₙ,, Yₙ ) → (X,Y) to show that the estimator converges to the true value. This property states that if sequences of random variables (Xₙ), (Yₙ) converge to X and Y respectively, then the joint sequence (Xₙ, Yₙ) converges to (X,Y).

User Skjcyber
by
8.6k points
Welcome to QAmmunity.org, where you can ask questions and receive answers from other members of our community.