Mini-project
Cybersecurity & Digital Trust
Phishing URL classifier
Predict whether a web address is a phishing site from features of the URL alone.
The brief
A browser extension needs to warn users about phishing links instantly, without visiting the page. Build classifiers that use only URL features, compare them, and test how well they hold up on newer phishing URLs. Good looks like: models compared fairly, a test on data from a different source or period, and a fast scoring function.
What you will build
- Engineer features from the URL string
- Train logistic regression, random forest and gradient boosting
- Test on a separate dataset or later period
- Measure speed per URL
- Package the best model as a scoring function
Free, open data to build it with
- PhiUSIIL phishing URL dataset ↗ UCI Machine Learning Repository
- Phishing Websites ↗ UCI Machine Learning Repository
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.