Your Cart

Your cart is empty

Home  >  Volume 28. No.2 (Nov. 2014)

26. Investigating the Effects Of Prior in Analyzing Reliability... by Iyiegbu T.I and Gulumbeand S.U - Volume28, No. 2, (November, 2014), pp 189 – 194
Sale price: $5.00


Analyzing Reliability problems often require the use of ‘Design of Experiment’ approach. The purpose of the optimal experimental design is to improve statistical inference regarding the quantity of interest by the optimal selection of values for design factors under the control of the investigator, within, of course the constraint of available resources. The standard practice of selecting a single model from some class of models and then making inference based on this model ignores model uncertainty. Ignoring model uncertainty can impair predictive performance and lead to overstatement of the strength of evidence via p-values that are too  small. Bayesian model averaging (BMA) provides a coherent approach for accounting for models uncertainty. Prior assumptions can be extremely critical for the outcome of BMA analyses. As uses and application of BMA increases with time, we wish to investigate how often prior assumptions may affect our inference as regards to improving reliability. Using data on reliability, we assess the effect of prior settings for posterior inference on model  size. We also consider the spread of model probabilities over model space, andthe posterior inclusion probabilities of the regressors.