Some of the data generated has issues.

[zkan12]

Thank you for your work for the open-source community! While running the code, (I chose the number of clusters to be 6) I noticed that the same dataset may produce different results, and the differences between the results are significant. I have also checked the problematic data, such as the ROI delineation, and found no issues. I wonder if this is due to the initial randomization in unsupervised clustering. If we want the data to be more reliable, how should we handle or select the data? I hope you can reply to this question despite your busy schedule. Wish you a smooth life.