Table 1. Overview of the classification tasks.
Classification Task Model Types and number of variables included Training data set data points (before oversampling) Training data set data points (after oversampling) Test data set data points
Injury
prediction		 linear SVM Training load, blood, fitness, questionnaire, jump variables (247) 1.085
(incl. 7 injuries)		 1.347
(incl. 269 injuries)		 300
(incl. 4 injuries)
Illness
prediction		 linear SVM Blood variables
(40)		 283
(incl. 11 illnesses)		 543
(incl. 271 illnesses)		 60
(incl. 1 illness)
Illness
determination		 linear SVM Blood variables
(40)		 177
(incl. 9 illnesses)		 337
(incl. 169 illnesses)		 42
(incl. 1 illness)
Injury prediction: 11 injuries present in the sample, with 8 different players affected; Illness prediction: 12 illnesses present in the sample, with 10 different players affected; Illness determination: 10 illnesses present in the sample, with 5 different players affected. Differences in the numbers of illness prediction and detection possible due to non-appearance for blood collection, when already ill.