I'm trying to build an Orange workflow to perform LDA topic modeling for analyzing a text corpus (.CSV dataset).

Unfortunately, the LDA widget in Orange lacks for advanced settings when comparing it with traditional coding in R or Python, which are commonly used for such purposes.

Accordingly, I would inquire about how to use Orange to:

  1. Measure (estimate) the optimal (best) number of topics ⁉️.

  2. Measuring topic-coherence score in LDA Topic Model in order to evaluate the quality of the extracted topics and their correlation relationships (if any) for extracting useful information ⁉️.

Is there a simple way that can accomplish these tasks in Orange ⁉️.

The following link provides the traditional solution for calculating the topic coherence score using Jupiter-Python as pre-explained ✅

Article Source link

I assembled the code-cells into a single file attached below:

Full Jupiter-Python Code:

Sample Dataset Corpus

Looking forward to your suggestions.

Thanks for your cooperation in advance.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.