I Taught a Computer to Spot Fake Websites With 96% Accuracy

Automated excerpt

Traditional defenses (blacklists of known bad URLs) are always playing catch-up. I used the UCI Phishing Websites dataset — 11,055 real websites, each labeled phishing or legitimate, described by 30 features. When I checked which features mattered most, SSL certificate state and anchor link behavior dominated — by a wide margin over the other 28 features.

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.