Table 6.

Macro accuracy results of the winning classifiers for each of the considered models.

Exercise type	Macro accuracy/winning classifier
	Total score	Component 1	Component 2	Component 3	Component 4	Component 5	Component 6
Sitting 1 and sitting 2	0.90/Gaussian process	0.88/Gaussian process	0.90/kNNa	0.89/Gaussian process	N/Ab	N/A	N/A
Sitting 3	0.87/Gaussian process	0.86/Neural network	0.91/Gaussian process	N/A	N/A	N/A	N/A
Standing 1 and standing 2	0.85/Gaussian process	0.83/Gaussian process	0.86/Gaussian process	N/A	N/A	N/A	N/A
Standing 3 (progressions 0-1)	0.91/kNN	0.91/Gaussian process	0.92/Gaussian process	0.89/kNN	0.90/Random forest	N/A	N/A
Standing 3 (progression 2)	0.87/SVMc (linear)	0.89/Gaussian process	0.90/Naïve Bayes	0.88/Random forest	0.91/kNN	N/A	N/A
Standing 3 (progression 3)	0.91/Random forest	0.90/AdaBoost	0.88/Neural network	0.86/kNN	0.89/kNN	N/A	N/A
Standing 4	0.92/Gaussian process	0.86/Gaussian process	0.88/Gaussian process	0.80/kNN	N/A	N/A	N/A
Walking 1	0.90/Random forest	0.81/Gaussian process	0.85/Random forest	0.92/Random forest	N/A	N/A	N/A
Walking 2 and walking 3	0.81/kNN	0.74/kNN	0.75/SVM (linear)	0.78/SVM (RBFd)	0.71/kNN	0.75/SVM (RBF)	0.75/kNN

^akNN: k-nearest neighbors.

^bN/A: not applicable.

^cSVM: support vector machine.

^dRBF: radial basis function.