TABLE 2.

The classification report of SVM classifiers.

		Number of observations	Cross- validation accuracy	Accuracy	Precision	Recall	F1-score	Matthews correlation coefficient
Naïve Bayes	Validation dataset
	Barefoot	53	0.89		0.89	0.91	0.90	0.78
	Shod	49			0.90	0.88	0.89
	Test dataset
	Barefoot	25		0.93	1.00	0.88	0.94	0.87
	Shod	19			0.86	1.00	0.93
SVM	Validation dataset
PCA-based SVM model	Barefoot	54	0.96		0.93	1.00	0.96	0.92
	Shod	48			1.00	0.92	0.96
RFE-based SVM model	Barefoot	51	0.98		0.98	0.98	0.98	0.96
	Shod	51			0.98	0.98	0.98
	Test dataset
PCA-based SVM model	Barefoot	24		0.95	0.96	0.96	0.96	0.91
	Shod	20			0.95	0.95	0.95
RFE-based SVM model	Barefoot	27		0.95	0.93	1.00	0.96	0.91
	Shod	17			1.00	0.88	0.94

Note: SVM, support vector machine; PCA, principal component analysis; RFE, recursive feature elimination.