cluster 数量很大的情况下如何处理？

[LeoCai]

cluster数量（比如category）很多的情况下会不会对耗时造成影响？ 另外cluster发生变化（比如新增category）的时候如何处理呢？

[stellaxu]

Hi, Leo Cai.

You post a very good question. Yes, when cluster number is very big (e.g. more than 500), the time consuming is a big issue. And when new category is added, then the model structure will be abruptly changed.

They are the issues to be handled later, and also welcome maybe new ideas from you to solve this issue, and maybe can contribute to a new research work and paper, haha.