(a) Mexican hat kernel and homogeneous clustered data

(b) Mexican hat kernel and heterogeneous clustered data

(c) Morlet kernel and homogeneous clustered data

(d) Morlet kernel and heterogeneous clustered data

Figure 1: Classification accuracy rates of the different types of feature extraction methods: PCA (blue plots), STIWK PCA (green plots), MTIWK PCA (red plots) and MDWK PCA (black plots), for clustered data. The green plots correspond to the mean value of the classification accuracy rates of STIWK which are calculated over all values of the parameter used in the STIWK.