- studied machine learning classification models (logistic regression, Naive Bayes , SVM) in more depth
- Learned how to evaluate a classification model via performance metrics and confusion matrices
- Grew familiar with hyperparameter tuning
- sci-kit learn
- trained four machine learning classification models, and optimized the hyper parameters via Gridsearch.
- Improved accuracy by 15-25% per model, the highest accuracy being 70% for logistic regression
- Understand BERT (at least at theoretical level) by reading online articles
- train BERT classifier on forum data