Material-based Stratification

MacQueen, D H

Lawrence Livermore National Laboratory

关键词: Probability;    99 General And Miscellaneous//Mathematics, Computing, And Information Science;    Stratification;    Sampling;    Detection;   

DOI  :  10.2172/919964
RP-ID  :  UCRL-TR-231455
RP-ID  :  W-7405-ENG-48
RP-ID  :  919964

美国|英语

来源: UNT Digital Library

【 摘 要 】

A simple probability model was applied to detection sampling in a room or space in which different surface materials are present. The model assesses the overall detection capability when the sampling and analytical methods have different performance properties for the different materials. The results suggest that some common sampling strategies may not be ideal. In particular: (1) In a single room or area that includes different surface types with different detection properties, do not use a single sampling grid with a common spacing throughout. (2) If it is known or strongly suspected that one material has better detection properties than the other, place all samples on that material. (3) When it is completely unknown which material has the better detection properties, allocate the samples equally between them.