Skip to main content

Advertisement

Table 3 Correlations for the different distance definitions

From: Classification complexity in myoelectric pattern recognition

  Average result Individual results  
  LDA (AMI) Single/OVO MLP AMI/MIX SVM AMI/MIX LDA (AMI) Single/OVO MLP AMI/MIX SVM AMI/MIX Data set
Mahalanobis 0.72/0.91 0.90/0.91 0.79/0.80 0.81/0.92 0.84/0.85 0.70/0.68 SM
0.78/0.88 0.86/NA 0.71/NA 0.85/0.91 0.80/NA 0.60/NA IM
Bhattacharyya 0.74/0.97 0.98/0.97 0.79/0.82 0.69/0.91 0.93/0.91 0.66/0.65 SM
0.83/0.96 0.96/NA 0.68/NA 0.79/0.89 0.94/NA 0.68/NA IM
Kullback–Leibler 0.60/0.88 0.93/0.90 0.65/0.70 0.54/0.76 0.84/0.82 0.63/0.60 SM
0.51/0.72 0.80/NA 0.32/NA 0.65/0.75 0.87/NA 0.65/NA IM
Hellinger 0.68/0.94 0.98/0.96 0.75/0.77 0.69/0.90 0.93/0.91 0.66/0.65 SM
0.80/0.95 0.97/NA 0.66/NA 0.79/0.89 0.94/NA 0.68/NA IM
Modified Mahalanobis 0.92/0.97 0.92/0.95 0.94/0.95 0.79/0.91 0.88/0.89 0.74/0.71 SM
0.93/0.94 0.87/NA 0.83/NA 0.85/0.90 0.86/NA 0.71/NA IM
  1. Correlations under “individual results” were calculated using classification accuracies and SIs from every individual movement, subject and feature, while those under “average result” were derived using the average SI and classification accuracy per subject and feature. Both methods provide one correlation, although “individual results” use more data. Classifiers were configured using AMI or MIX. Classifiers were used in the conventional “single” topology, apart from LDA, which was used in “single” and OVO. The highest correlation values per column are highlighted in bold. All correlations were found to be statistically significant at p < 0.01. The MIX configuration is not applicable (NA) for individual movements since there is not mixed outputs