Table 2 Comparison of the model’s performance showing the median AUC over 20 repeats of nested cross-validation, using all data modalities and each modality separately
From: Predicting dementia in people with Parkinson’s disease
Modality | Random forest | XGBoost | Logistic regression | |||
|---|---|---|---|---|---|---|
Median AUC | P-value | Median AUC | P-value | Median AUC | P-value | |
Multimodal | 0.62 (0.005) | – | 0.60 (0.009) | – | 0.62 (0.006) | – |
Demographics + Genetics | 0.61 (0.004) | <0.05 | 0.59 (0.007) | <0.05 | 0.61 (0.006) | <0.05 |
Demographics | 0.56 (0.005) | <0.05 | 0.54 (0.008) | <0.05 | 0.57 (0.004) | <0.05 |
Genetics | 0.58 (0.003) | <0.05 | 0.56 (0.008) | <0.05 | 0.58 (0.05) | <0.05 |
Comorbidities | 0.58 (0.004) | <0.05 | 0.57 (0.004) | <0.05 | 0.57 (0.03) | <0.05 |
Lifestyle | 0.49 (0.01) | <0.05 | 0.5 (0.009) | <0.05 | 0.49 (0.005) | <0.05 |
Environmental | 0.51 (0.007) | <0.05 | 0.51 (0.008) | <0.05 | 0.5 (0.006) | <0.05 |
Family history | 0.49 (0.01) | <0.05 | 0.49 (0.01) | <0.05 | 0.49 (0.006) | <0.05 |
