Table 11.
Overview of selected comparable studies.
| Condition | References | Sample size | Type | Algorithm | Accuracy | Precision | Recall | AUROC |
|---|---|---|---|---|---|---|---|---|
| ADHD | Garcia-Argibay et al. (39) | 238,696 | CS | DL | 69% | NR | 72% | 0.75 |
| ADHD | Maniruzzaman et al. (40) | 45,779 | CS | RF | 86% | NR | 86% | 0.94 |
| ADHD | Ter-Minassian et al. (41) | 4,178 | CS | LR | 81% | 70% | 66% | 0.72 |
| Disruptive behaviors | Menon and Krishnamurthy (42) | 1,100 | CS | CNN | 72% | NR | 70% | 0.74 |
| CD | Chan et al. (14) | 2,368 | Long | FNN | 91% | NR | 93% | 0.96 |
Summary of comparable studies using machine learning to predict externalizing disorders in youth that are reviewed in this manuscript. Numbers after the authors' names correspond to the citation reference. CS, cross-sectional; Long, longitudinal; DL, deep learning; RF, random forest; LR, logistic regression; CNN, convolutional neural network; FNN, feedforward neural network.