Project / 02
Hybrid ML-Based Phishing URL Detection
A hybrid machine learning system combining URL lexical and structural features with DNS and SSL reputation analysis for real-time phishing URL classification.
Category
Machine Learning · Cybersecurity
Key Result
99% Accuracy
The project
A hybrid phishing detection system designed to classify URLs by combining machine learning with domain reputation signals. The approach uses URL lexical and structural features together with DNS and SSL reputation analysis.
System architecture
URL
Lexical / Structural Features
DNS Reputation
SSL Reputation
XGBoost
Classification
Technical details
ML CLASSIFIER
XGBoost
Used as the machine learning model for phishing URL classification.
DOMAIN REPUTATION
DNS
DNS reputation signals are incorporated into the hybrid detection approach.
DOMAIN REPUTATION
SSL
SSL reputation signals provide an additional domain-level signal.
Measured performance.
99%
Accuracy
0.99
Precision
0.99
Recall
0.99
F1 Score
1.00
ROC-AUC
<1%
False Positive
Next Project