Benchmarking Machine Learning Uncertainty Quantification Methodologies for Predicting Turbine Gas Temperature Degradation

In three linesBenchmarking of 5 uncertainty quantification methods (Delta, Bayesian Monte Carlo Dropout, Bootstrap, LUBE, MVE) for turbine gas temperature degradation prediction. Evaluation on real dataset using coverage probability and prediction interval width metrics. Trade-offs identified between accuracy and reliability.

Summary generated by Claude — human-verified