Fig. 1

Stacked ensemble learning using several fitness-fatigue and ML models. Briefly, let \({\mathbf{X}}_{m \times n}\) be a first level training data set of \(m\) features and n observations. Predictions made from the models constitute a prediction matrix (i.e. a second level training data set) of dimension \(n \times P\) base-models. Finally, a combiner—or a meta-model—denoted \({\mathcal{M}}\) is trained on these data to predict the final outcome \(\hat{y}^{{\text{f}}}\)