wvangansbeke/Unsupervised-Classification

Accuracy becomes lower after self-labeling

kkellyk opened this issue · 0 comments

Hi author,
I got an accuracy of 0.66 using my dataset scan,
but after self-labeling it turned out to be 0.33.
Why are self-labeling results worse?
Am I doing something wrong?
How can this situation be improved?

Please author to help me answer
Thanks
Kelly