Replicate accuracy results from benchmark datasets

[stevcabello]

Hi,

I'm wondering how can I replicate these results (the ones in bold for the Earthquakes, WormsTwoClass and Strawberry datasets):

I have tried using the benchmark_test.py in this way, for example--> python3 benchmark_test.py --dataset ucr-WormsTwoClass (same for other datasets), but results are different than the ones from the figure.

I'm aware there are several more parameters that I could use in the command above, but would like to know which ones from the different parameters found in benchmark_test.py should I use:

Thanks!