Algorithm [Ref.] | Dataset | Number of classes | Methods | Accuracy |
---|---|---|---|---|
SVM [20] | MITBPD | 2 | HRV, DFA, windowed DFA | 79.31 ± 4.52% for 12 features 79.99 ± 4.64% for 10 features |
RF [21] | SHRSV | 3 | HRV | 88.67% for subject specific classifier 72.58% for subject independent classifier |
ELM [22] | MITBPD | 2 | HRV and DFA | 78.33% for both HRV and DFA; 76.29% for HRV; 73.48% for DFA |
BPNN [22] | MITBPD | 2 | HRV and DFA | 76.74% for both HRV and DFA; 73.81% for HRV; 71.07% for DFA |
SVM [22] | MITBPD | 2 | HRV and DFA | 78.12% for both HRV and DFA; 76.03% for HRV; 73.21% for DFA |
SVM with ECOC extension and RBF [25] | MITBPD | 2 and 3 | HRV and EDR | 81.76% for 2 classes 76% for 3 classes |
LSTM network [23] | Siesta | 4 | HRV | 77.00 ± 8.90% |
Combination of SVM and PSO [24] | MITBPD | 2 | HRV | 78.41% for RBF kernel 77.08% for linear kernel |
Morphological [29] | MITBPD | 3 | HRV | 77.02% |
Deep Neural Network [30] | MITBPD | 3 | HRV | 77% |
Combination of ELM and PSO (Proposed Method) | MITBPD | 2 | HRV | 81.96% for 2 classes 76.59% for 3 classes 71.44% for 4 classes 63.18% for 6 classes |