Contribution to Reduce Type II Error at the End-of-Line Inspection at an Automotive Industry Supplier
Contribution to Reduce Type II Error at the End-of-Line Inspection at an Automotive Industry Supplier
Purpose – In the automotive industry, it is necessary to create improvement opportunities to increase customer satisfaction and to stand out from the competition. Components to be assembled in a car are subject to a final end-of-line (EOL) inspection, which are measured using numerical and attributive variables, with a final result of OK / NOK. However, this inspection is not perfect. In this paper, a method is proposed to increase the detection rate of products with components in a pre-failure state, which pass the EOL inspection.
Design/methodology/approach – The logistic regression method made it possible to identify an explanatory model, based on real data from a company of the automotive industry, that allows the product to be categorised into one of two groups: “0km” (customer claim) or “Regular” (conforming product).
Findings – It was possible to explore the Type II error reduction through the usage of logistic regression.
Originality/value – The logistic regression proved to be valuable in classifying at EOL potential prefailure devices, having the potential to reduce the Type II error. It also identifies the key variables that support the classification.
100% End-of-line inspection, Logistic regression, Pre-failure, Type II error
