Learned vs. hand-designed features for ECG beat classification: A comprehensive study

Keyphrases

• Learned Features 100%
• Handcrafted Features 100%
• ECG Beat Classification 100%
• Multilayer Perceptron 66%
• Support Vector Machine 66%
• 1D Convolutional Neural Network 66%
• Deep Neural Network Classifier 66%
• Performance Improvement 33%
• Computational Efficiency 33%
• Classification Performance 33%
• Single Channel 33%
• Data-centric 33%
• ECG Data 33%
• High Classification Accuracy 33%
• Principal Coordinate Analysis (PCoA) 33%
• Benchmark Dataset 33%
• MIT-BIH Arrhythmia Database 33%
• Convolutional Neural Network 33%
• Convolutional Layer 33%
• Hermite Transform 33%
• Dyadic Wavelet Transform 33%
• ECG Classification 33%

Computer Science

• Learned Feature 100%
• Convolutional Neural Network 100%
• Multilayer Perceptron 66%
• Support Vector Machine 66%
• Performance Improvement 33%
• Experimental Result 33%
• Comparative Evaluation 33%
• Classification Performance 33%
• Classification Accuracy 33%
• Computational Efficiency 33%
• Component Analysis 33%
• Principal Components 33%
• Convolutional Layer 33%
• Dyadic Wavelet Transform 33%