Home
10K Spanish scam urgency phrases (4 countries) → Production ML training
✅ 1st parallel Spanish-English scam corpus (Multi-dialect)
✅ 97.8/100 ML readiness (zero noise, perfect labels)
✅ Psycholinguistic urgency triggers (fear/panic/greed)
✅ Regional dialect authenticity
✅ 12 ML features → BERTspa F1: 92-94%
📁 Sample-Scammer-Phrases-ByCountry_100rows.csv
✅ 100% Complete (zero nulls) | 100% UTF-8 BOM | 94% Class Balance
✅ 8 countries × 5 sentiments × 7 scam types = Perfect stratification
✅ Native validation: Mexico(98%)/Spain(97%)/Argentina(96%)
✅ vs Real data: +18% F1 improvement | 0% label noise
[🏠 Home] | 🔬 ML Quality | 🌍 Dialects | 📊 Features | 🚀 Production | 💼 Business
Next: ML-Quality