A Copula-based Multivariate Degradation Analysis for Reliability Prediction

Degradation test often involves multivariate Performance Characteristics (PCs) to be analyzed to make reliability prediction. As a result, the complex dependency structure among PCs needs to be addressed. In this paper, we develop a flexible copula-based multivariate model for analyzing highdimensional degradation process. A two-stage method for parameters estimation is developed as an efficient statistical inference scheme. Finally, a real LED dataset is analyzed by the proposed approach.