New concept to compute confidence of reported information level for logic diagnosis
Résumé
This paper proposes a model to compute con dence of reported information level (CRIL) in the domain of logic diagnosis. This level of con dence is provided by a diagnosis module allowing to quickly identify the origin of equipment failure. We studied the factors a ecting CRIL, such as measurement system reliability, production context, position of sensors in the acquisition chains, type of product, reference metrology, preventive maintenance and corrective maintenance based on historical data and reported information generated by production equipment. We have introduced a new 'CRIL' concept based on the Bayesian Network approach, Na ve Bayes model and Tree Augmented Na ve Bayes model. Our contribution includes an on-line con dence computation module for production equipment data, and an algorithm to compute CRIL. We suggest it be applied to the semiconductor manufacturing industry.
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...