Table 9 Comparative analysis of the proposed model and the existing state-of-the-art methods.
Reference | Dataset used | Method used | Accuracy (%) | Comparison with the proposed model |
|---|---|---|---|---|
MIT-BIH | Bi-LSTM with an attention mechanism | 87.84 | Lower | |
MRI T1 Maps | Deep learning for hypertrophic cardiomyopathy | 89.79 | Lower | |
ECG signals | Machine learning for congenital heart disease | 90.65 | Lower | |
Tabular data | Heuristic-metaheuristic ensemble model | 91.34 | Lower | |
Various datasets | Self-attention transformer model | 90.71 | Lower | |
IoT-enabled ECG data | Three-layer deep learning with meta-heuristic methods | 92.41 | Lower | |
ECG signals | Convolutional block attention network | 90.03 | Lower | |
Audio signals | AI-based classification of audio signals | 89.00 | Lower | |
Pediatric ECG | Deep learning with human concepts | 89.79 | Lower | |
General cardiovascular data | Machine learning-based predictive models | 91.52 | Lower | |
Clinical ECG data | Artificial neural networks and sensors | 93.02 | Lower | |
Echocardiography data | AI for rheumatic heart disease detection | N/A | Not applicable | |
Various datasets | Classification with Boruta feature selection | 91.00 | Lower | |
Multi-modal data | Deep learning with multi-modal data fusion | 90.85 | Lower | |
ECG data | Various deep learning techniques | N/A | Not applicable | |
ECG data | Deep learning for ECG signal detection | 91.02 | Lower | |
Chinese diabetic population | Machine learning for coronary heart disease | 91.80 | Lower | |
Various datasets | Deep learning with crowd intelligence optimization | N/A | Not applicable | |
Proposed multiclass model | ECG signals from MIT and INCART | Modified Multiclass Attention Mechanism with Deep BiLSTM | 98.82 | Higher |
