psegedy / antispam

Spam filter demonstration using machine learning

antispam

Spam filter demonstration using machine learning

Classify emails in .eml format if it is SPAM or HAM using scikit-learn python module. Before first run it is neccessary to have training dataset for the classifier. Simply change DATASET list with path to training data and values True for SPAM and False for HAM. For this demonstration I'm using some emails from this pre-classified dataset http://www2.aueb.gr/users/ion/data/enron-spam/

Dependencies

scikit-learn html2text scipy pandas numpy

Example of usage

./antispam.py email.eml email2.eml email3.eml email4.eml

Output

email.eml - OK 
email2.eml - SPAM
email3.eml - SPAM 
email4.eml - FAIL

• OK - it's HAM
• SPAM - it's SPAM
• FAIL - failed to load or classify email

Note: Once classifier is trained, you can comment call to train_data() method. It is project for BIS class @FIT VUT and it was submitted with trained classifier.