In reliability analyses, the primary objective is to model the failure times of the product under study. In many situations, these failures occur because the product degrades (weakens) over time. But, sometimes failures do not occur. In these situations, modeling the product degradation over time is helpful in making predictions about failure times. When an accelerating factor is included in the data, you can fit an accelerated degradation model.
The Degradation platform can model data that follows linear or nonlinear degradation paths. If a path is nonlinear, transformations are available to linearize the path. If it is not possible to linearize the path, you can specify a nonlinear model.
You can also use the Degradation platform to perform stability analysis. Three types of linear models are fit, and an expiration date is estimated. Stability analysis is used in setting pharmaceutical product expiration dates. Stability models that use pooled mean square error (MSE) are also supported.