Key result
XGBoost achieves 99% heart sound classification accuracy, outperforming hybrid CNN+LSTM models via feature engineering.
Why the study?
Early and accurate heart sound-based detection of cardiovascular diseases is critical in clinical diagnostics to improve patient outcomes.
Does training machine learning models (XGBoost) on engineered acoustic features improve heart sound classification accuracy compared to Wav2Vec embeddings?
Does training machine learning models (XGBoost) on engineered acoustic features improve heart sound classification accuracy compared to Wav2Vec embeddings?
Domain-specific feature engineering combined with XGBoost provides highly accurate (99%) multi-class classification of heart sounds, outperforming transformer-based embeddings.
Engineered features may aid research pipelines for heart sound analysis; extends prior work on domain-specific ML but leaves clinical adoption open.
Heart sound-based detection of cardiovascular diseases is a critical task in clinical diagnostics, where early and accurate identification can significantly improve patient outcomes. In this study, we investigate the effectiveness of combining traditional acoustic features and transformer-based Wav2Vec embeddings with advanced machine learning models for multi-class classification of five heart sound categories. Ten engineered acoustic features, i.e., Log Mel, MFCC, delta, delta-delta, chroma, discrete wavelet transform, zero-crossing rate, energy, spectral centroid, and temporal flatness, were extracted as regular features. Four model configurations were evaluated: a hybrid CNN + LSTM and XGBoost trained with either regular features or Wav2Vec embeddings. Models were assessed using a held-out test set with hyperparameter tuning and cross-validation. Results demonstrate that models trained on regular features consistently outperform Wav2Vec-based models, with XGBoost achieving the highest accuracy of 99%, surpassing the hybrid model at 98%. These findings highlight the importance of domain-specific feature engineering and the effectiveness of ensemble learning with XGBoost for robust and accurate heart sound classification, offering a promising approach for early detection and intervention in cardiovascular diseases.
No takes yet. Share an insight, caveat, or question.
Karthikeya et al. (2026) studied Cardiovascular diseases. XGBoost with engineered acoustic features vs. Hybrid CNN + LSTM and Wav2Vec embeddings was evaluated on Multi-class classification accuracy of five heart sound categories. XGBoost trained on engineered acoustic features achieved 99% accuracy in classifying heart sounds, outperforming a hybrid CNN+LSTM model (98%) and Wav2Vec-based models.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: