Original source: http://www.dt.fee.unicamp.br/~tiago/smsspamcollection/
This trains a classifier using scikit-learn to label messages as 'ham' or 'spam'.
Training is optional, as the model files are already stored under the model/ subfolder.
Download Anaconda, select the Python 3.x version.
conda create -n simple-ml python=3
conda activate simple-ml
conda install pandas jupyter scikit-learn
jupyter notebook
Download dataset from the link above, and extract the files into the data/ sub-folder.
Run through simple-sms-spam-classifier.ipynb. If all goes well, the model files under the model/ subfolder will be updated.
Requires: NodeJS v8 or later
# Run from the simple-ml anaconda environment
npm install
npm start
- Open browser to http://localhost:3000
- Input a SMS string (up to 160 characters)
- Click 'Spam or Ham?'
