Jay Qadan
Reputation: 87
How to detect topics in arbitrary text file or data? not knowing number of topics beforehand
I have textual data that I want to discover topics it has, I used trained doc2vec on large corpus such as Wikipedia, but there is inconsistency in the results. Is there a better approach to discover topics.
Upvotes: 0
Views: 254
Answers (1)
WolfgangK
Reputation: 993
You could try using Hierarchical Dirichlet Process as implemented in the gensim class gensim.models.hdpmodel.HdpModel. From the gensim docs:
Unlike its finite counterpart, latent Dirichlet allocation, the HDP topic model infers the number of topics from the data.
Upvotes: 1
Related Questions
- Topic modeling on short texts Python
- Detect predefined topics in a text
- how to label topics automatically after applying LDA
- Extract Topic Scores for Documents LDA Gensim Python
- Topic distribution: How do we see which document belong to which topic after doing LDA in python
- infer topic distributions on new, unseen documents with LDA and Gensim
- Calculating topic distribution of an unseen document on GenSim
- Text Processing, How to assign 1 topic -> 1 document using LDA?
- LDA for Topic Modeling in Python
- Finding topics of an unseen document via Gensim