search

maximtrp / bitermplus

Biterm Topic Model (BTM): modeling topics in short texts
https://bitermplus.readthedocs.io/en/stable/
MIT License
77 stars 13 forks source link

Topics' names? #36

Closed ldbfufu closed 1 year ago

ldbfufu commented 1 year ago

I wonder where the names of the topics exist?

maximtrp commented 1 year ago

What do you mean by names? Topic1, topic2, etc? You can create such labels yourself. Just convert a phi or theta matrix to DataFrame. Or you can use tmplot helper methods for this.

ldbfufu commented 1 year ago

Thank you for your response. Like, I want to know what is the topic of the first document. Software? Can the model automatically give the corresponding topic name for each document? Which file is the phi or theta parameter located in?

maximtrp commented 1 year ago

Topic modeling cannot do it. Topic labels are usually assigned manually by a group of researchers after the analysis of top topic words and documents. This is a difficult and time consuming task. tmplot package I created was meant to help users create various tables with top words and documents, which should be used for labeling.