Ocena reprezentatywności pobierania próbek
Abstract
THE PROBLEM OF TAKING REPRESENTATIVE SAMPIES
Summary
The paper deals with the problem of taking representative samples in the case when any feature of a given series analysed changes continuously. To obtain a representative picture of the whole series we must take samples within the series analysed, according to a uniform density net, i.e. we must take care to obtain samples that represent similar volume. Under such conditions, when the amount of sampling points are determined, any sample must representatively determine the features of a given series within an area (having radius r). On this basis a possibility exists to determine average weight of a sample. The weight should always be proportional to the area with the radius r determined previously. Average value of an error ε, calculated on the basis of samples having known probability level, may also be determined. The amount of samples characterized by various weight is a quotient of the entire area and of the elementary areas with the radius r.
Summary
The paper deals with the problem of taking representative samples in the case when any feature of a given series analysed changes continuously. To obtain a representative picture of the whole series we must take samples within the series analysed, according to a uniform density net, i.e. we must take care to obtain samples that represent similar volume. Under such conditions, when the amount of sampling points are determined, any sample must representatively determine the features of a given series within an area (having radius r). On this basis a possibility exists to determine average weight of a sample. The weight should always be proportional to the area with the radius r determined previously. Average value of an error ε, calculated on the basis of samples having known probability level, may also be determined. The amount of samples characterized by various weight is a quotient of the entire area and of the elementary areas with the radius r.