Change-Point Problem With Correlated Observations, With an Application in Material Accountancy

Abstract

Techniques are developed for the change-point problem when the data are a series of correlated normal variables, combining locally most powerful testing with Bayesian estimation. The approach is applied to a nuclear materials accounting problem. The effect of correlation on change-point tests is examined.

KEY WORDS:

• Correlation
• Locally most powerful testing
• Bayesian estimation
• Nuclear safeguards