Generalization over accuracy: A cross-dataset, explainable, and federated learning framework for Parkinson's disease detection.

Effect size analysis enables the following insights: • Identification of meaningful improvements beyond random variation • Differentiation between statistically noticeable and clinically relevant gains • Improved interpretability of performance changes across experimental conditions This approach ensures that reported improvements are not only statistically observable but also practically significant for real-world deployment. 3.10.4.

also in 1,558 other papers