Utilization of the error entropy value as a criterion of the precision of instruments and measurements
If it is acknowledged unquestionably that the object of measuring devices consists of obtaining measurement information, it is necessary to select as a basic index of the quality of these devices a criterion which has a singlevalued relationship to the measured information content. Such a criterion consists of the entropy error value Δ of a measuring device or a measurement.
For a sufficiently large number of tests it is possible to determine from relationship (5) the precise value of Δ for the random error component. For a small number of tests an approximate value of Δ can be obtained from Δ=K · σ. In such a case the systematic errors can also be taken into account if σ is calculated as the root-mean-square deviation from the nominal value.
The historically established and legally confirmed procedure for certifying measuring instruments can be considered as one of the simplest approximate methods for a statistical determination of precisely the error entropy values. A further improvement of these rules may consist of establishing a more correct method for determining the error entropy value of instruments and measurements.
The main and basic significance of the error entropy concept consists not so much in improving the rules for certifying instruments, but rather in providing a more precise basic concept of the theory of measurements by determining that which characterizes their precision from the information-theory point of view. Without rejecting the existing certification methods this concept helps to find the degree of their approximation to the basic characteristic which consists of the error entropy value.
KeywordsEntropy Systematic Error Information Content Basic Concept Random Error
Unable to display preview. Download preview PDF.