Accession Number : ADA136890

Title :   Inspections with Unknown Detection Probabilities.

Descriptive Note : Technical rept.,

Corporate Author : STANFORD UNIV CA

Personal Author(s) : Derman,C ; Lieberman,G J ; Ross,S M

PDF Url : ADA136890

Report Date : Nov 1983

Pagination or Media Count : 24

Abstract : This document assumes that in an acceptance sampling situation the lot is subject to 100% inspection. The probability that a defective unit is detected is different for each inspector and is unknown. It is of interest to estimate N, the number of defective units in the lot (presumably, a decision to reject or accept the lot would be based on the estimate of N). Or, suppose satellites are used for surveillance over a given part of the earth with the detection of certain types of installations being the mission of a given satellite. However, for various reasons, it can be assumed the detection of any existing installation is uncertain with an unknown probability of detection that varies among satellites. The problem is to estimate the total number of installtions based on the number observed. A third situation involves the reading of a manuscript by many proofreaders. Based on the results, it may be of interest to estimate the total number of typographical errors.

Descriptors :   *Mathematical models, *Error analysis, *Maximum likelihood estimation, *Typography, Iterations, Statistical distributions, Convergence, Random variables

Subject Categories : Statistics and Probability
      Printing and Graphic Arts

Distribution Statement : APPROVED FOR PUBLIC RELEASE