En jämförelse av stickprovsmetoder vid mätning av radioaktivetet på en yta
SemivariogramRandom fieldSimple random samplingSpatial samplingSystematic samplingStatisticsOperations researchProgrammingActuarial mathematicsStatistikOperationsanalysProgrammeringAktuariematematikMathematics and Statistics
This study compares the sample variance of the mean in a simple random sample (SRS) and a systematic sample from a surface. The study also compares the prediction error variance in a SRS and systematic sample from a random field. Recommendations are presented on which method to use in the clearance and exemption process in the nuclear industry. The results indicate that quadratic systematic sampling outperformed SRS in most instances and proved to be especially efficient in the presence of a long-ranged covariance function and high sampling intensity (i.e. short distances between observations). Due to the lack of efficient estimators, one should use the same estimator as with SRS when estimating the sample variance from a systematic sample.