Implementing Naïve Bayes Classifier, LSTM and BERT on

SPAM/HAM and 20 News Group

SPAM/HAM
- Easy and quick
- Bag of words model
- 97.8% Accuracy
20 News Group
- Easy and quick for a larger dataset as well
- Bag of words model for focused observations
- 72.8% Accuracy

SPAM/HAM
- Not so dense architecture
- 87.15% accuracy
20 News Group
- Simple architecture considering the size of the dataset
- Validation accuracy not more than 33.5%
- While high training accuracy, we are left with very low validation accuracy

SPAM/HAM
- Using pretrained BERT model with Transformers
- Validation Accuracy at 98.21%
20 News Group
- Using pretrained BERT model with Transformers
- Validation Accuracy at 68.2%

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
20 Mini News Group BERT.ipynb		20 Mini News Group BERT.ipynb
20 News NBC LSTM.ipynb		20 News NBC LSTM.ipynb
README.md		README.md
Spam Ham NBC LSTM BERT.ipynb		Spam Ham NBC LSTM BERT.ipynb
Spam Ham SVC NBC RF.ipynb		Spam Ham SVC NBC RF.ipynb