Bias (Statistics)

Note

Bias is the difference between the expected value [=mean] of an estimator and the ground truth.

$$\text{Bias}[\hat{\theta }]=\mathbb{E}[\hat{\theta }]-{\theta }_0$$

thm. Linear Transformation Preserves Bias. If $\hat{\theta }$ is unbiased, $g(\hat{\theta })$ is unbiased if $g$ is a linear transformation. → For estimators without bias, we can compute the precision to evaluate how good it is.