Bias, Variance and Mean Squared Error
While bias quantifies the average difference to be expected between an estimator and an underlying parameter, an estimator based on a finite sample can additionally be expected to differ from the parameter due to the randomness in the sample.
One measure which is used to try to reflect both types of difference is the expected mean square error,
This can be shown to be equal to the square of the expected bias, plus the expected variance:
![\begin{align}
\operatorname{MSE}(\hat{\theta})= & (\operatorname{E})^2 + \operatorname{E})^2\,]\\
= & (\operatorname{Bias}(\hat{\theta},\theta))^2 + \operatorname{Var}(\hat{\theta})
\end{align}](http://upload.wikimedia.org/math/4/d/0/4d067a1ff288751d4e66d504684916ac.png)
An estimator that minimises the bias will not necessarily minimise the expected mean square error.
Read more about this topic: Bias Of An Estimator
Famous quotes containing the words variance, squared and/or error:
“There is an untroubled harmony in everything, a full consonance in nature; only in our illusory freedom do we feel at variance with it.”
—Fyodor Tyutchev (1803–1873)
“I want that head so sanitary and squared away that the Virgin Mary herself would be proud to go in there and take a dump.”
—Stanley Kubrick (b. 1928)
“You can hardly convince a man of an error in a life-time, but must content yourself with the reflection that the progress of science is slow. If he is not convinced, his grandchildren may be. The geologists tell us that it took one hundred years to prove that fossils are organic, and one hundred and fifty more to prove that they are not to be referred to the Noachian deluge.”
—Henry David Thoreau (1817–1862)