Application used to classify whether an SMS message is likely a spam message or ham (legitimate) message.
It tests various classification and vectorization technique combinations to determine which combination works best on the SMS Spam dataset provided from Kaggle.
Run classifier.py
to view the scores of each of the combinations.
One-Vs-Rest Classification and Term Frequency-Inverse Document Frequency vectorization were chosen to perform the determinations.
Application makes classifications on SMS messages with a 98.8% accuracy rate, due to the aforementioned ML techniques above.
- Python
- Flask
- scikit-learn (Machine Learning library)
- pandas (Data Analysis and Manipulation library)
- SQLite
- SQLAlchemy
- HTML
- CSS
Run the below:
mkdir sms-spam-detector-python
cd sms-spam-detector-python
git clone https://github.com/malcolmrichardson/sms-spam-detector-python.git
virtualenv venv
source venv/bin/activate
pip install -r requirements.txt
python app.py
Application runs on https://127.0.0.1:5000 by default.
Enjoy and thank you!