Text Classification Chula course description

This repository is a part of the subject: 2147334 MACHINE LEARNING OR DEEP LEARNING

Getting start:

Google Colab:
Link
Jupyter notebook:
docker-compose -f docker-compose.dev.yml up jupyter --build
Jupyter notebook with pytorch GPU support (Take longer to load):
docker-compose -f docker-compose.dev.yml up jupyter-pytorch --build

Objectives:

1. Predict study programs by looking from course description
2. (Optional) Use embedding from the best predictor to calculate course similarity and in the end create content-based recommender system.

Idea:

• Get courses data and study programfrom Academic Chula website
• There are about 500 study programs, 20,000 coursesm and 26 faculties in Chulalongkorn university.
• Use model that support course description embedding ex. TF-IDF, Neural network based model.
• Use SHAP to interpret models' result

Roadmap:

• Train model to predict study programs and compare between TF-IDF + SVM, LSTM + Embedding layer, and Transformer based model. Have very ppor result - Early stopped
  • LSTM
  • TF-IDF + SVM
  • Thai2Fit + SVM
  • Attention based model
• Add model notebook.
• Add scrapper notebook.
• Get study program data.
• Get courses description data.
• Preprocess courses description data.
• Train model to predict faculty
  • LSTM
  • TF-IDF + SVM
  • Thai2Fit + SVM
  • Pretrained Attention based model (WangChanBERTA)
• Create recommender system proof of concept from trained model