- python3
- sklearn (machine learning models)
- joblib ( save the models on disk)
See text-classification.ipynb for a classification using Naive Bayes.
Note: Hyper-parameters are not tuned for any of the models. Moreover accuracy depends on type of features selected, train-test set partition overfitting etc.
- all the features ( more than 1000K) were taken for all models, As expected KNN performed very poorly.