Symptom trajectory classification on CES-D personalized threshold label (test)
62Balanced AccuracyXGBoost
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| XGBoostModel Type=Machine Learning2026.04 | 62 | 75 | 63.4 | 16.6 | |
| SVMModel Type=Machine Learning2026.04 | 62 | 75.1 | 61 | 19.1 | |
| ElasticNetModel Type=Machine Learning, Highlight=Best AUC2026.04 | 60.3 | 75.5 | 56.1 | 17.5 | |
| LightGBMModel Type=Machine Learning2026.04 | 58.6 | 73.1 | 46.3 | 14.5 | |
| Regression to person's meanModel Type=Rule-based baseline, Highlight=Strongest baseline2026.04 | 53.3 | 64.2 | 29.3 | 24 | |
| Person-specific modal classModel Type=Rule-based baseline2026.04 | 33.3 | 50 | 0 | — | |
| Predict all stableModel Type=Rule-based baseline2026.04 | 33.3 | 50 | 0 | — | |
| Last value carried forwardModel Type=Rule-based baseline2026.04 | 32.2 | 52.9 | 4.9 | 4.5 |