Mit Bih Ecg Class Eval

mit-bih-ecg-class-eval

Deep Neural Network Architectures for Electrocardiogram Classification: A Comprehensive Evaluation — Yun Song et al. (2026) (arXiv:2602.17701, 2026)

What this evaluates

Evaluates deep learning architectures for multi-class ECG heartbeat classification, specifically testing their ability to handle severe class imbalance and capture temporal cardiac waveform features.

Datasets

• MIT-BIH Arrhythmia Database — total ?; splits: train (-1), val (-1), test (-1)

Metrics

• F1-score (primary) — range: [0, 1]
  • Harmonic mean of Precision and Recall: 2 * (Precision * Recall) / (Precision + Recall). Used as the headline metric to balance sensitivity and specificity under class imbalance.
• Accuracy — range: [0, 1]
  • Proportion of correctly classified samples out of the total number of samples.
• Macro-average AUC — range: [0, 1]
  • Average of the Area Under the Receiver Operating Characteristic Curve computed independently for each arrhythmia class, ensuring equal weighting across all categories.

Input / output format

Input: Preprocessed ECG waveform segments (heartbeats) representing temporal cardiac signals.

Output: Predicted arrhythmia class label (e.g., Normal, Atrial Premature, Fusion of ventricular and normal).

Scoring recipe

def compute_f1(y_true, y_pred):
    tp = sum(1 for t, p in zip(y_true, y_pred) if t == p)
    precision = tp / sum(1 for p in y_pred if p == 1)
    recall = tp / sum(1 for t in y_true if t == 1)
    return 2 * precision * recall / (precision + recall) if (precision + recall) > 0 else 0.0

Common pitfalls

• Class imbalance heavily skews Accuracy and Precision; F1-score or macro-averaged metrics must be prioritized for fair comparison.
• Macro-average AUC requires per-class ROC curves and probability outputs, not just hard predictions.
• Confusion matrix off-diagonal elements often reveal morphologically similar classes (e.g., Normal vs. Atrial Premature) that are frequently misclassified.

Evidence (verbatim from paper)

To quantify classification performance, the study adopted standard evaluation metrics, including Accuracy, Precision, Recall, and F1-score, along with Receiver Operating Characteristic (ROC) and Area Under the Curve (AUC) analyses. Accuracy measures the overall proportion of correctly classified samples, while Precision and Recall capture the trade-off between false positives and false negatives. The F1-score, defined as the harmonic mean of Precision and Recall, provides a balanced measure of sensitivity and specificity, making it particularly suitable for the imbalanced class distribution characteristic of ECG datasets.

Citation

@misc{song2026deepneural,
  title={Deep Neural Network Architectures for Electrocardiogram Classification: A Comprehensive Evaluation},
  author={Yun Song et al. (2026)},
  year={2026},
  note={arXiv:2602.17701}
}

• arXiv: 2602.17701