Skip to main content

Table 5 Sensitivity, specificity, and accuracy of combining various feature extraction methods with various classifiers to distinguishing between Set E (ictal windows) and Sets (A, D) (normal and interictals).

From: Combination of EEG Complexity and Spectral Analysis for Epilepsy Diagnosis and Seizure Detection

Window size Classifier Feature Selection Sensitivity Specificity Accuracy
173 LLS All features 93.61 (2.1) 97.79 (0.8) 96.40 (0.6)
   GA 92.85 (2.5) 97.69 (0.8) 96.08 (0.7)
   ApEn + AR model 92.53 (1.8) 98.10 (0.7) 96.25 (0.7)
  LDA All features 95.60 (1.4) 97.32 (0.9) 96.74 (0.5)
   GA 95.67 (1.7) 96.99 (1.2) 96.55 (0.7)
   ApEn + AR model 94.03 (1.7) 97.55 (0.8) 96.38 (0.7)
  BP All features 94.59 (2.5) 97.42 (1.2) 96.47 (0.8)
   PCA 94.26 (2.3) 97.91 (0.8) 96.70 (0.8)
   GA 89.34 (3.9) 98.45 (0.5) 96.63 (0.5)
   ApEn + AR model 93.97 (2.4) 97.38 (1.1) 96.24 (0.7)
  LISVM All features 94.75 (1.8) 97.84 (0.9) 96.81 (0.5)
   PCA 93.91 (2.2) 97.64 (0.8) 96.40 (0.6)
   GA 93.97 (2.6) 97.62 (1.0) 96.40 (0.9)
   ApEn + AR model 94.36 (2.1) 97.42 (1.0) 96.40 (0.7)
  RBFSVM All features 95.22 (1.9) 97.98 (0.9) 97.06 (0.6)
   PCA 98.85 (1.9) 97.90 (0.8) 96.88 (0.6)
   GA 94.75 (2.2) 97.72 (1.1) 96.73 (0.8)
   ApEn + AR model l 94.07 (1.8) 97.23 (1.1) 96.18 (0.6)
256 LLS All features 94.92 (1.5) 98.37 (0.8) 97.22 (0.5)
   GA 94.33 (1.7) 98.15 (0.7) 96.88 (0.4)
   ApEn + AR model 93.39 (1.8) 98.08 (0.8) 96.49 (0.7)
  LDA All features 96.06 (1.2) 97.77 (1.0) 97.20 (0.5)
   GA 95.92 (1.8) 97.24 (1.3) 96.80 (0.6)
   ApEn + AR model 96.36 (1.7) 97.51 (0.9) 96.46 (0.7)
  BP All features 95.89 (2.1) 97.79 (1.4) 97.16 (0.9)
   PCA 95.25 (1.5) 98.58 (0.7) 97.47 (0.5)
   GA 95.75 (3.0) 97.35 (1.4) 96.82 (1.1)
   ApEn + AR model 94.16 (2.6) 97.80 (1.2) 96.58 (0.6)
  LISVM All features 96.19 (1.8) 98.23 (1.0) 97.55 (0.6)
   PCA 95.08 (1.9) 98.20 (0.9) 97.16 (0.5)
   GA 95.45 (2.6) 97.91 (1.2) 97.09 (0.9)
   ApEn + AR model 94.83 (2.0) 97.62 (1.0) 96.69 (0.6)
  RBFSVM All features 96.88 (1.8) 98.41 (1.0) 97.90 (0.7)
   PCA 96.67 (1.7) 98.43 (1.0) 97.83 (0.6)
   GA 96.06 (2.5) 98.17 (0.9) 97.47 (0.7)
   ApEn + AR model 93.80 (2.2) 97.09 (1.2) 95.99 (1.7)
512 LLS All features 95.59 (1.3) 98.58 (0.7) 97.58 (0.3)
   GA 95.50 (2.0) 98.39 (1.0) 97.43 (0.9)
   ApEn + AR model 93.94 (2.0) 98.13 (0.8) 96.73 (0.7)
  LDA All features 97.03 (1.0) 98.13 (0.9) 97.76 (0.4)
   GA 96.81 (1.7) 97.78 (0.9) 97.46 (0.9)
   ApEn + AR model 94.59 (1.9) 97.77 (0.9) 96.71 (0.7)
  BP All features 97.44 (1.5) 97.81 (1.4) 97.69 (0.9)
   PCA 96.78 (1.6) 98.47 (0.9) 97.90 (0.5)
   GA 95.91 (2.7) 98.05 (1.3) 97.33 (1.0)
   ApEn + AR model 95.00 (1.8) 97.88 (1.0) 96.92 (0.6)
  LISVM All features 97.28 (1.4) 98.66 (0.8) 98.20 (0.5)
   PCA 96.38 (1.3) 98.47 (0.9) 97.77 (0.5)
   GA 96.88 (2.0) 98.33 (1.1) 97.84 (0.7)
   ApEn + AR model 94.69 (2.5) 97.70 (0.9) 96.70 (0.8)
  RBFSVM All features 97.53 (1.4) 98.73 (1.0) 98.33 (0.7)
   PCA 97.44 (1.5) 98.77 (0.9) 98.32 (0.6)
   GA 97.37 (1.9) 98.53 (0.9) 98.15 (0.8)
   ApEn + AR model 93.63 (2.4) 97.28 (1.3) 96.06 (1.0)