Table 5 Comparison between different window lengths for tremor prediction based on EEG data processing
From: Parkinson’s disease tremor prediction using EEG data analysis-A preliminary and feasibility study
Feature set | Classifier | Accuracy (%) | Sensitivity | Specificity | AUC | |||||
|---|---|---|---|---|---|---|---|---|---|---|
Feature set No.1 | 3-s | KNN | 73.67 ± 9.56 | 70.90 ± 14.89 | 87.18 ± 10.00 | 61.52 ± 14.53 | 90.54 ± 8.92 | 92.50 ± 6.10 | 79.4 ± 9.30 | 0.74 ± 0.09 |
5-s | Nave Bayes | 73.82 ± 18.78 | 78.61 ± 24.47 | 88.45 ± 15.04 | 69.90 ± 29.16 | 92.80 ± 10.45 | 85.58 ± 14.05 | 86.00 ± 13.36 | 0.75 ± 0.18 | |
Feature set No.2 | 3-s | KNN | 81.29 ± 9.19 | 79.15 ± 11.38 | 84.86 ± 11.53 | 83.94 ± 20.93 | 90.37 ± 10.33 | 95.62 ± 6.29 | 86.57 ± 4.50 | 0.81 ± 0.09 |
5-s | Nave Bayes | 66.70 ± 17.00 | 66.58 ± 27.93 | 74.96 ± 25.17 | 65.39 ± 21.25 | 90.86 ± 10.72 | 85.69 ± 13.97 | 75.02 ± 19.88 | 0.67 ± 0.17 | |