Abstract
The diagnosis of acceptance or rejection of a lot by sampling inspection can be treated as a pattern recognition problem. In this paper some contributions of the pattern recognition methodology to design of sampling plan are described and illustrated. After a brief description of pattern recognition especially the Bayes' diagnosis rule, a procedure is presented for design of single sampling inspection by attributes. The recommended procedure enables us to find an optimal sampling plan which insures the specified consumer's risk (or producer's risk) and minimizes the average loss due to misclassification of a lot.
