A0175
Title: On reliability analysis of one-shot device testing data with defects
Authors: Man Ho Ling - The Education University of Hong Kong (Hong Kong) [presenting]
Abstract: The problem of defective devices in the manufacturing industry is considered. Defective devices can arise for various reasons, such as errors made by workers, inadequate quality processes, insufficient training, or issues with reliability during the design stage. The focus is on one-shot device test data that includes defects that have occurred during a realistic manufacturing process. In this scenario, the question arises of whether a failed device is defective or has failed due to its lifetime being shorter than the inspection time. The maximum likelihood approach is explored to estimate the mean-time-to-failure based on a sample of one-shot devices with manufacturing defects. This will be done using gamma and Weibull lifetime distributions. We will also examine how masking can affect the estimation and analysis under different defective rates and masking proportions. In addition, a Bayesian approach is presented to deal with cases with low defective rates.
