iNeuron
Automated machine learning classification pipeline detecting malicious and spoofed domains in real time.
- Extracted lexical, WHOIS, and network-based features from over 50,000 domain URLs.
- Trained and evaluated ensemble tree-based models and logistic regression baselines.
- Tuned classification thresholds to systematically minimize false positive rates.
- Achieved 90.3% accuracy in domain threat classification.
- Reduced false positives by 12%, preventing unnecessary alert fatigue for security operators.