Table 3.
Accuracy of activity recognition algorithms.
| Accuracy | Data | Stacking | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Learning Algorithms | Classifiers | (No Temperature) | 0% | 10% | 20% | 30% | 50% | 70% | 90% |
| SVM | 74% | 81% | 40% | 43% | 43% | 48% | 47% | -% | |
| KNN | 73% | 81% | 80% | 80% | 83% | 83% | 85% | 93% | |
| Conventional Machine Learning | SGD | 50% | 47% | 52% | 55% | 53% | 55% | 50% | 54% |
| LR | 54% | 56% | 58% | 56% | 58% | 56% | 55% | 58% | |
| NB | 59% | 61% | 63% | 61% | 62% | 62% | 62% | 59% | |
| Deep Learning | SDAE | 75% | 77% | 69% | 72% | 74% | 75% | 75% | 86% |
| RF | 78% | 88% | 89% | 89% | 89% | 89% | 92% | 91% | |
| Ensemble Learning | ET | 78% | 89% | 88% | 89% | 89% | 90% | 92% | 96% |
| DF | 65% | 71% | 77% | 76% | 74% | 75% | 75% | 79% |