Skip to content
Snippets Groups Projects
Select Git revision
  • main default protected
  • FinalSubmission
2 results
rawal's avatar
Update README.md
rawal authored
3999de96
History
rawal's avatar 3999de96
History
Name Last commit Last update
modified
original
.gitignore
Cosine Similarity-LDA.jpeg
Cosine Similarity-PCA.jpeg
README.md
Report.pdf
Topic_Modeling.ipynb
Variance-LDA.jpeg
Variance-PCA.jpeg
helper.py
main.py
metrics.py
requirement.txt
README.md

PGM Exam 2022

Getting started

You can run the code from

  1. Notebook: Provided in this repo or in colab using:

https://colab.research.google.com/drive/1khqHZc-m3qx2ZO7XjsRc3KeHhOuS2gTV?usp=sharing

  1. Python file
python main.py

Results

The results were generated using colab with tfidf_features in function main set to 10000 and it took about 30 minutes. For fast execution you may set it to 2000, hence it is set as default. But setting it to 2000 will only give result for cosine similarity in PCA for LDA it has to be higher.

Results where the topics are identified between original and modified dataset:

Failed experiments results where topics are not identified between original and modified dataset: