Professional Documents
Culture Documents
The theory of estimation deals with the assessment of unknown parameters of the general
population that are based on a random experiment.
1.3.1 Estimator
A basic term of the theory of estimation is an estimator. This is understood as a clearly
defined function of the outcomes of a random trial implied by an unknown parameter θ of
the general population. If so, the estimator is a random variable.
Notice that if the point of interest is an unknown parameter θ of the general population
and that a random sample of size n is taken to estimate its value, then by applying the estima-
tor Tn we have the following situation:
tn f ( x , x2 , ..., xn ) (1.89)
d = Tn – tn (1.90)
The right side of the equation determines a certain random variable because the function
of a random variable is also a random variable. Therefore, the left side of the equation is also
a random variable. Therefore, the estimation error is a random variable. It has its own prob-
ability distribution and we are able analyse its basic parameters.
To achieve a precise estimation, i.e. to ensure a small error of estimation, it is necessary
to pay attention to both the correct sampling and the selection of an estimator with good
statistical properties.
A good estimator should be, among the other things, characterised by following properties:
• unbiasedness
• consistency
• efficiency
• sufficiency.