Is this topic model suitable for modeling chinese corpus?
Sorry for the late reply! I'd definitely say the library is usable for Chinese. You will, however need a Chinese tokenizer to be able to segments texts in a meaningful way. You can, for instance use the jieba library. Otherwise I see no obstacle that would prevent you from applying it to Chinese texts.
Sorry for the late reply! I'd definitely say the library is usable for Chinese. You will, however need a Chinese tokenizer to be able to segments texts in a meaningful way. You can, for instance use the jieba library. Otherwise I see no obstacle that would prevent you from applying it to Chinese texts.
Thanks for your patient explanation,I will consider using this topic-modeling in my future research.
I have, in fact, written a whole Medium article about topic modelling in Chinese, you should consider reading it :D It uses a novel topic model of ours, called KeyNMF and uses representations from transformer models.
That's great. I will learn about this model. It's a great honour to learn about your research results.
发自我的iPad
------------------ Original ------------------ From: Márton Kardos @.> Date: Sat,Mar 15,2025 10:54 PM To: centre-for-humanities-computing/tweetopic @.> Cc: Chonglin Pan @.>, Author @.> Subject: Re: [centre-for-humanities-computing/tweetopic] Is this topic modelsuitable for modeling chinese corpus? (Issue #22)
I have, in fact, written a whole Medium article about topic modelling in Chinese, you should consider reading it :D It uses a novel topic model of ours, called KeyNMF and uses representations from transformer models.
Contextual Topic Modelling in Chinese Copora with KeyNMF
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***> x-tabdeveloping left a comment (centre-for-humanities-computing/tweetopic#22)
I have, in fact, written a whole Medium article about topic modelling in Chinese, you should consider reading it :D It uses a novel topic model of ours, called KeyNMF and uses representations from transformer models.
Contextual Topic Modelling in Chinese Copora with KeyNMF
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>