Heart disease detection -model comparison
Led a 5-person team comparing five ML classification algorithms on a clinical dataset. Rigorous evaluation prioritized real-world viability (precision, recall, AUC) over simplistic accuracy metrics.
Algorithms compared
5
Team size
5 members
Metrics tracked
precision · recall · AUC
Dataset
healthcare (clinical)
Rigorous model evaluation
Compared decision trees, SVMs, neural networks, and ensemble methods. Emphasis on clinical viability: a high-accuracy model that misses true positives is worse than a lower-accuracy model with better recall.