Fig. 9.

Classification accuracy for the best sets corresponding to distance definitions of SI. Boxplot of average classification accuracy over all movements when using the best sets representing the distance definitions found in the legends. The middle line of the box is the median, the marker is the mean, and the box extends to the 25th and the 75th percentiles for the bottom and the top, respectively. The different insets compare sets of different number of features. The result is derived from the IM data set. Classifiers were used in the conventional “single” topology, apart from LDA, which was used in “single” and OVO