An Application of Bayesian Methods for Combining Data from Different Test Modalities
This report documents the research into the application of hierarchical Bayesian methods for characterizing the population failure rate (i.e. probability of defect) of an electronic component based on test data from a number of different test modalities. Classical statistical methods, those based on a frequency approach permit the combination of point estimates but stumble when characterizing the resulting confidence limits. Classical Bayesian methods permit the logical combination of test data, but are not fully efficient in incorporating all available information. In particular, classical Bayesian methods assume that the articles under test are not related in any manner even though the articles may be identical. Alternatively, hierarchical Bayesian methods permit the relationship between test articles to be explicitly included in the analysis. Data from four different test modalities are considered in the analysis. Comparisons are made between the current analysis approach (using traditional statistical methods), classical Bayesian methods and a hierarchical Bayesian approach.
- Research Organization:
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States); Sandia National Lab. (SNL-CA), Livermore, CA (United States)
- Sponsoring Organization:
- US Department of Energy (US)
- DOE Contract Number:
- AC04-94AL85000
- OSTI ID:
- 809604
- Report Number(s):
- SAND2002-3953; TRN: US200307%%640
- Resource Relation:
- Other Information: PBD: 1 Mar 2003
- Country of Publication:
- United States
- Language:
- English
Similar Records
Data from: Comparing traditional and Bayesian approaches to ecological meta-analysis
Evaluating the Benefits of Bayesian Hierarchical Methods for Analyzing Heterogeneous Environmental Datasets: A Case Study of Marine Organic Carbon Fluxes