evaluated accuracy seems weird

[leihouyeung]

Hi, thanks for your great work. I had an issue that when I used my own data set to conduct fully unsupervised clustering. The outputted accuracy showed a random choice. For example, I have 3 classes and the numbers of these data are balanced. But the accuracies in 5 sub_heads were all around 0.333. Did I set the parameters wrong? Could you help me with this? Thanks a lot!