Title :
Comparison of classical and Bayesian models using BLR
Author :
Mense, A.T. ; Olsen, Grant
Author_Institution :
Raytheon Co., Tucson, AZ, USA
Abstract :
The modeling of storage reliability using binary logistic regression allows for one to explore the causal or correlative variable relationships that affect the outcome of binary (pass/fail) tests. Both classical and Bayesian models show good agreement when the amount of data available is large as it should. Bayesian models should add some value for situations in which there are only small amounts of testing information.
Keywords :
Bayes methods; regression analysis; reliability; storage management; BLR; Bayesian models; binary logistic regression; binary test; causal variable relationships; classical models; correlative variable relationships; pass/fail tests; storage reliability modeling; Bayes methods; Data models; Equations; Logistics; Mathematical model; Missiles; Reliability; Bayesian; logistic regression; storage reliability;
Conference_Titel :
Reliability and Maintainability Symposium (RAMS), 2014 Annual
Conference_Location :
Colorado Springs, CO
Print_ISBN :
978-1-4799-2847-7
DOI :
10.1109/RAMS.2014.6798479
