Fig. 4: Comparison of different clustering options.

a Shows the difference between using the pure manometry values and applying the change filter. Using the pure manometry values leads to missing a few rare, yet important classes, which can be seen when applying the change filter. b Shows the comparison of different clustering methods. Agglomerative clustering and k-means achieve similarly distinctive clusters that include all important classes, while DTW-based k-means results in slightly less distinctive clusters.