From: EEG-based classification of epilepsy and PNES: EEG microstate and functional brain network features
 | All | All-alpha | All-beta | All-delta | All-theta |
---|---|---|---|---|---|
Accuracy, Precision, Recall using occurrence (k) as the discriminative (input) feature | |||||
 Random Forest | 0.696, 0.732, 0.696 | 0.700, 0.738, 0.700 | 0.536, 0.537, 0.536 | 0.708, 0.741, 0.708 | 0.734, 0.784, 0.734 |
 SVM (Linear) | 0.732, 0.758, 0.732 | 0.734, 0.768, 0.734 | 0.396, 0.391, 0.396 | 0.724, 0.756, 0.724 | 0.700, 0.725, 0.700 |
 SVM (RBF) | 0.730, 0.774, 0.730 | 0.724, 0.772, 0.724 | 0.524, 0.535, 0.524 | 0.728, 0.772, 0.728 | 0.758, 0.793, 0.758 |
 Decision Tree | 0.646, 0.683, 0.646 | 0.680, 0.711, 0.680 | 0.602, 0.615, 0.602 | 0.656, 0.682, 0.656 | 0.680, 0.715, 0.680 |
 kNN | 0.794, 0.814, 0.794 | 0.822, 0.841, 0.822 | 0.694, 0.708, 0.694 | 0.762, 0.788, 0.762 | 0.802, 0.831, 0.802 |
 Gradient Boost | 0.692, 0.729, 0.692 | 0.696, 0.746, 0.696 | 0.654, 0.669, 0.654 | 0.688, 0.732, 0.688 | 0.698, 0.731, 0.698 |
Accuracy, Precision, Recall using duration (k) as the discriminative (input) feature | |||||
 Random Forest | 0.686, 0.735, 0.686 | 0.684, 0.741, 0.684 | 0.538, 0.544, 0.538 | 0.674, 0.747, 0.674 | 0.670, 0.731, 0.670 |
 SVM (Linear) | 0.690, 0.740, 0.690 | 0.740, 0.806, 0.740 | 0.464, 0.401, 0.464 | 0.614, 0.668, 0.614 | 0.662, 0.733, 0.662 |
 SVM (RBF) | 0.614, 0.758, 0.614 | 0.590, 0.713, 0.590 | 0.598, 0.623, 0.598 | 0.694, 0.716, 0.694 | 0.578, 0.678, 0.578 |
 Decision Tree | 0.662, 0.685, 0.662 | 0.718, 0.743, 0.718 | 0.584, 0.587, 0.584 | 0.730, 0.758, 0.730 | 0.644, 0.668, 0.644 |
 kNN | 0.724, 0.746, 0.724 | 0.744, 0.773, 0.744 | 0.530, 0.541, 0.530 | 0.776, 0.820, 0.776 | 0.740, 0.765, 0.740 |
 Gradient Boost | 0.754, 0.796, 0.754 | 0.772, 0.810, 0.772 | 0.632, 0.636, 0.632 | 0.722, 0.767, 0.722 | 0.710, 0.743, 0.710 |
Accuracy, Precision, Recall using coverage (k) as the discriminative (input) feature | |||||
 Random Forest | 0.730, 0.776, 0.730 | 0.734, 0.782, 0.734 | 0.560, 0.567, 0.560 | 0.710, 0.769, 0.710 | 0.754, 0.794, 0.754 |
 SVM (Linear) | 0.674, 0.702, 0.674 | 0.664, 0.696, 0.664 | 0.460, 0.439, 0.460 | 0.672, 0.707, 0.672 | 0.672, 0.704, 0.672 |
 SVM (RBF) | 0.618, 0.6631, 0.618 | 0.620, 0.666, 0.620 | 0.380, 0.354, 0.380 | 0.634, 0.681, 0.634 | 0.616, 0.656, 0.616 |
 Decision Tree | 0.698, 0.711, 0.698 | 0.738, 0.753, 0.738 | 0.682, 0.692, 0.682 | 0.678, 0.704, 0.678 | 0.628, 0.661, 0.628 |
 kNN | 0.776, 0.806, 0.776 | 0.764, 0.786, 0.764 | 0.634, 0.644, 0.634 | 0.778, 0.802, 0.778 | 0.738, 0.784, 0.738 |
 Gradient Boost | 0.762, 0.804, 0.762 | 0.782, 0.815, 0.782 | 0.696, 0.707, 0.696 | 0.758, 0.797, 0.758 | 0.748, 0.792, 0.748 |
Accuracy, Precision, Recall using combination of occurrence (k), duration (k) and coverage (k) as the discriminative (input) feature | |||||
 Random Forest | 0.678, 0.738, 0.678 | 0.698, 0.764, 0.698 | 0.562, 0.567, 0.562 | 0.688, 0.757, 0.688 | 0.710, 0.779, 0.710 |
 SVM (Linear) | 0.712, 0.753, 0.712 | 0.734, 0.792, 0.734 | 0.514, 0.488, 0.514 | 0.612, 0.634, 0.612 | 0.656, 0.742, 0.656 |
 SVM (RBF) | 0.530, 0.337, 0.530 | 0.500, 0.250, 0.500 | 0.494, 0.308, 0.494 | 0.514, 0.393, 0.514 | 0.508, 0.268, 0.508 |
 Decision Tree | 0.692, 0.725, 0.692 | 0.698, 0.719, 0.698 | 0.612, 0.624, 0.612 | 0.660, 0.690, 0.660 | 0.646, 0.678, 0.646 |
 kNN | 0.724, 0.746, 0.724 | 0.744, 0.773, 0.744 | 0.530, 0.540, 0.530 | 0.776, 0.820, 0.776 | 0.740, 0.7652, 0.740 |
 Gradient Boost | 0.754, 0.792, 0.754 | 0.772, 0.805, 0.772 | 0.702, 0.712, 0.702 | 0.714, 0.763, 0.714 | 0.758, 0.786, 0.758 |