Consistency|Satistical Inefernece



Q. What is Consistency?

Ans: An estimator $\small T_{n}= T(x_{1},x_{2},....x_{n})$ based on random sample of size $\small n$ is said to be consistent estimator of $\small \gamma(\theta)$, where $\small \theta$ belongs to the $\small \Theta$, if $T_{n}\overset{p}{\rightarrow}\gamma(\theta)$ as $\small n \to \infty$. 

$\small P{|T_{n}-\gamma(\theta)|<\varepsilon}\rightarrow 1$ as $\small n \rightarrow \infty$

Example: If $\small T_{n}'$ = $\small (\frac{n-a}{n-b})T_{n}$ and $\small T_{n}$ is a consistent estimator of $\small \gamma(\theta)$, then $\small T_{n}'$ is also consistent for $\small \gamma(\theta)$ as $\small n \rightarrow \infty$.

Theorem 1:

If $\small T_{n}$ is a consistent estimator of $\small \gamma (\theta)$ and $\small \varphi {\gamma(\theta)}$ is a cont. function of $\gamma (\theta)$, then $\small \varphi(T_{n})$ is a consistent estimator of $\small \varphi{\gamma(\theta)}$.

Theorem 2:

Sufficient conditions for Consistency

An estimator $\small T_{n}= T(x_{1},x_{2},....x_{n})$ based on random sample of size $\small n$ is said to be consistent estimator of $\small \gamma(\theta)$, where $\small \theta$ belongs to the $\small \Theta$, if

i. $\small E(T_{n}) \rightarrow \gamma(\theta)$ as $\small n \rightarrow \infty$

ii. $\ Var_{\theta}(T_{n})\rightarrow 0$ as $\small n \rightarrow \infty$

Example: 

Q.If $\small X_{1},X_{2},....,X_{n}$ are random observations on a Bernoulli variate $\small X$ taking the value $\small 1$ with probability $\small p$ and the value $\small 0$ with probability $\small (1-p)$ show that: 

$\frac{\sum x_{i}}{n}(1-\frac{\sum x_{i}}{n})$ is a consistnet estimator of $\small p(1-p)$

Solution: 

Since $\small X_{1},X_{2},.....,X_{n}$ are i.i.d Bernoulli variates with parameter 'p'

$\small T=\sum x_{i}\sim B(n,p)$ 

or,$\small E(T)=np$, and $\small V(T)=npq$

$\small \bar X=\frac{1}{n}\sum x_{i}=\frac{T}{n}$

or,$\small E(\bar X)=\frac{E(T)}{n}=\frac{np}{n}$

$\small V( \bar X)==\frac{V(T)}{n^2}=\frac{npq}{n^2}=0$ when $\small n\rightarrow \infty$

Since $\small E(\bar X)\rightarrow p$ and $\small V(\bar X)\rightarrow 0$ as$\small n\rightarrow \infty$. $\small \bar X$ is a consistent estimator of $\small p$

$\small \bar X(1-\bar X)$ being a polynomial of $\small \bar X$, is a continuous of p.

$\small \bar X$ is a consistent estimator of $\small p$.


Tags: consistency meaning in bengali, example of consistency, consistency statistics example,  consistency statistics formula, consistency in statistics pdf, how to calculate consistecy, consistency estimator pdf, unbiased but inconsistent estimator, how to calculate consistency


 


 

Post a Comment (0)
Previous Post Next Post