*4.3. Comparison Results of Classification Algorithms*

In this section, we use *p*<sup>1</sup> through *p*<sup>12</sup> to statistically compare the performance of different classifiers for subject AB01. The objective is to find the best classifier for locomotion mode detection among linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machine (SVM) with linear kernels (SVM-linear), SVM with RBF kernels (SVM-RBF), multi-layer perceptron (MLP), and decision tree (DT). The tuned parameter value of RBF kernel is *σ* = 1. Table 6 shows mean classification accuracy and standard deviation of each classifier trained with the features from each Pareto point using 10-fold cross validation (CV).


**Table 6.** Mean classification accuracy (ACC) and standard deviation (STD) for AB01 of classifiers trained with 13 different feature subsets. NF is the number of features in each set.

Table 7 presents pairwise statistical comparisons using Wilcoxon signed-rank tests at a 5% significance level. If a pairwise *p*-value is less than 0.05, the mean performances of the two classifiers are statistically significantly different, and the classifier with larger mean prediction accuracy performs better than the other one. A pairwise *p*-value greater than 0.05 indicates no significant difference between the performance of the two classifiers. Table 7 shows that the classification performance of MLP and SVM-RBF are statistically equal, and are significantly better than the other methods. SVM-linear is statistically better than LDA, QDA, and DT. QDA performs better than LDA and similarly to DT. In summary, MLP and SVM-RBF are the best, SVM-linear is the second best, QDA and DT are the third best, and LDA is the worst for locomotion mode detection.

**Table 7.** Comparison of classification performance using Wilcoxon signed-rank tests (W.T.) at a 5% significance level. B or W indicates that the row method performs better or worse than the column method, respectively, while T shows that they tie with similar performance. ∗ indicates that the lower triangular half of the table is equal to its upper triangular half. These results are obtained using all the data from Table 6.

