RQ2: Scalability for long time-series

Are the methods capable of solving the problem in a reasonable amount of time for a relatively long time series ?

To answer this question, we measure the influence of the length of the time series on the execution time of the methods. Our evaluation is summarized in the plot below.

Summary of the results

RQ2 Conclusion

Grammarviz, LatentMotif and Motiflets are the most scalable methods in terms of computational efficiency for handling long time series. STOMP, PanMP, MDL-Clust, PEPA, and A-PEPA have acceptable execution times up to 500,000 samples, while the other methods crash or exceed the timeout before that.