About Mean shift

[jieWANGforwork]

Thanks for your work, I have some quesitions about the code.
Could you provide the code of comparison methods or is there any way you can explain the following concerns:
Did you pretrain the cluster model then used the embeddings to evaluate all methods on unexpectiveness or implemented all methods with individual mean shift?
There was no clarification in the paper if I didn't miss any details.

That would be greatly helpful for myself to learn from your resaech.
Thanks for you patience.

[lpworld]

Hello! The Mean Shift algorithm is trained together with the recommendation model on the fly. You can refer to the codes from line 184-203.

[jieWANGforwork]

Thanks for your reply, and can you also tell me how to evaluate the unxpec of comparison methods? Thank you.