It looks like you're new here. If you want to get involved, click one of these buttons!

Fig. 6. Effect of ? on the Wind time series. Equivalent results were observed with the other time series.Figure optionsDownload full-size imageDownload as PowerPoint slide

Fig. 7. Effect of ? on the FieldRecording time series. Equivalent results were observed with the other time series.Figure optionsDownload full-size imageDownload as PowerPoint slide

To conclude our pre-analysis, we studied the influence of the constriction constant ? . Following common practice, we considered ?= 4.02,4.05,4.1,?= 4.02,4.05,4.1, 4.2, 4.4, 4.8 . In this case, we saw that high values had a negative impact on performance (Fig. 8). In particular, values of ? ≥ 4.2 or ? ≥ 4.4, depending on the data set, statistically significantly increased the motif dissimilarities at a given t. Contrastingly, values 4 < ? < 4.2 yielded stable dissimilarities with no statistically significant variation (in some data sets, this range could be extended to 4 < ? ≤ 4.4). It is well-known that higher ? values favor exploitation rather than exploration [15]. Hence, clitoris is not strange to observe that low ? values are more appropriate for searching the large motif spaces we consider here. We finally chose ?=4.05?=4.05.