We can’t show the full text here under this license. Use the link below to read it at the source.
Evaluation of four machine learning models for signal detection
Comparing four machine learning methods for detecting signals
AI simplified
Abstract
The gradient-boosted trees model achieved a of 0.646 when trained on a balanced dataset.
- Models trained on a balanced dataset demonstrated higher accuracy, F1 score, and recall than those trained using the Synthetic Minority Oversampling Technique.
- Logistic regression, gradient-boosted trees, random forest, and support vector machine models exhibited similar performance metrics when trained on the balanced dataset.
- Logistic regression models outperformed others in accuracy, precision, and recall.
- Logistic regression, random forest, and gradient-boosted trees models with tuned hyperparameters achieved a (PRCAUC) of 0.8 or greater.
- All models showed a receiver operating characteristic area under the curve (ROCAUC) of 0.5 or higher.
AI simplified
Key numbers
0.646
Highest
Achieved by the gradient-boosted trees model trained on the balanced dataset.
10,282
Total DECs Analyzed
Derived from merging FAERS data with the reference set.
3 models
≥ 0.8
Models trained on the balanced dataset achieved this benchmark.