Bayesian Two-Process Weibull Modelling of PCL-Blend-PEG Coating Degradation in Simulated Body Fluid: Composition-Held-Out Validation, Identifiability Limits and Exploratory Electrochemical Correlation

Description

Vlad-Cristian Stoica and Anita Ioana Visan model how implant coatings dissolve in body-like fluid, so sixteen-week soaking tests can be shortened. Their 2026 paper reanalyzes mass-loss curves for four polymer films on titanium, blends of a slow-degrading plastic with a fast-dissolving one. Published data offer only a handful of recipes, too few for greedy machine learning.

Their Bayesian model splits decay into two Weibull processes, fast dissolution and slow breakdown, with timing that depends on the blend. A simpler one-process model serves as the baseline under a preset decision rule. The authors also stress-test whether the data can even identify the richer model's parts, an honesty check modelers often skip.

Catalog record

Source: https://europepmc.org/articles/PMC13611490