Table 3.
Feature sets used as an input vector to produce the six models (in bold the features selected by CART algorithm) and their classification performance measures
| Model Features | CA | F1 | Precision | Recall | |
|---|---|---|---|---|---|
| ModelStw | MeanRR, SDNN, LFn, FD, sex, age | 60.2% | 58.6% | 57.7% | 60.1% |
| ModelCorr | MeanRR, SDNN, LF, LF/HF, Beta exp, sex, age | 61.4% | 59.1% | 58.1% | 61.4% |
| ModelAll_features | MeanRR, SDNN, RMSSD, NN50, pNN50, LF, HF, LF/HF, LFn, HFn, Beta exp, SD1, SD2, SD1/SD2, FD, sex, age | 60.3% | 58.2% | 57.2% | 60.3% |
| ModelStw+LVEF | MeanRR, pNN50, LF/HF, FD, sex,age, LVEF | 73.3% | 71.3% | 70.8% | 72.9% |
| ModelCorr+LVEF | MeanRR, SDNN, LF, LF/HF, Beta exp, sex, age, LVEF | 72.8% | 71.3% | 70.8% | 72.7% |
| ModelAll_features+LVEF | MeanRR, SDNN, RMSSD, NN50, pNN50, LF, HF, LF/HF, LFn, HFn, Beta exp, SD1, SD2, SD1/SD2, FD, sex, age, LVEF | 72.6% | 70.8% | 70.3% | 72.6% |