consistent estimator
Given a set of samples from a given probabilitydistribution with an unknown parameter , where is the parameter space that is a subset of .Let be an estimator of . Allowingthe sample size to vary, we get a sequence of estimators for:
We say that the sequence of estimators consistent (or that is a consistent estimator of), if converges in probability to forevery . That is, for every ,
for all .
Remark. Suppose is an estimator of such thatthe sequence is consistent. If and are two convergent sequences of constants with and , then the sequence , defined by , is consistent, is an estimator of .
Proof.
First, observe that
This implies
As , ,, and .So the last expression goes to as . Therefore,
and thus is a consistent sequence of estimatorsof .∎