Table 7.
Performance of the machine learning classifiers for public Fitbit datasets.
| ML classifier models | Mean accuracy | Precision | Recall | F1 | MCC |
|---|---|---|---|---|---|
| SVC (kernel = 'linear') | 90.00 | 90.00 | 90.00 | 90.00 | 87.16 |
| SVC (kernel = 'rbf') | 92.14 | 92.00 | 92.00 | 92.00 | 90.03 |
| GaussianNB () | 81.78 | 82.00 | 82.00 | 82.00 | 76.82 |
| DecisionTreeClassifier (criterion = "gini") | 97.10 | 97.00 | 97.00 | 97.00 | 96.10 |
| DecisionTreeClassifier (criterion = "entropy") | 97.50 | 97.00 | 98.00 | 97.00 | 96.78 |
| RandomForestClassifier (n_estimators = 50) | 95.99 | 96.00 | 96.00 | 96.00 | 95.29 |
| RandomForestClassifier (n_estimators = 100) | 95.99 | 96.00 | 96.00 | 96.00 | 95.40 |
| KNeighborsClassifier (n_neighbors = 2) | 89.40 | 89.00 | 89.00 | 89.00 | 86.10 |
| KNeighborsClassifier (n_neighbors = 4) | 92.50 | 92.00 | 92.00 | 92.00 | 90.36 |
| DummyClassifier | 33.21 | 11.00 | 33.00 | 11.00 | 00.00 |