| Modul | Algoritm | Asosiy Ko'rsatkich | Qiymat | Izoh |
|---|---|---|---|---|
| Anomaliya | Isolation Forest | F1-Score | 0.83 | Tree-based, n_estimators=200 |
| Anomaliya | One-Class SVM | F1-Score | 0.78 | RBF kernel, nu=0.05 |
| Anomaliya | Local Outlier Factor | F1-Score | 0.78 | n_neighbors=5, density-based |
| Anomaliya | ๐ Ensemble (Majority) | F1-Score | 0.89 | 3 algoritm birlashmasi โ ENG YAXSHI |
| Klaster | K-means (k=5) | Silhouette Score | 0.63 | n_init=20, 5 ta foydalanuvchi guruhi |
| Klaster | K-means (k=5) | Davies-Bouldin | 1.42 | Pastroq = yaxshiroq ajratish |
| DBSCAN | DBSCAN (eps=0.8) | Noise % | ~8% | Avtomatik anomal foydalanuvchilar |
| Performance | Flask + pandas | API Throughput | 2.86 GB/min | Preprocessing tezligi |
| UX | SUS Questionnaire | SUS Score | 78.5 | 15 foydalanuvchi ยท "Good to Excellent" |
| Algoritm | Precision | Recall | F1-Score | Vizual | Aniqlangan | Tavsif |
|---|
| Ko'rsatkich | Natija | Vizual | Izoh |
|---|---|---|---|
| Preprocessing Throughput | 2.86 GB/min | 500K+ log yozuvi qayta ishlash | |
| API Median Response | 120ms | REST API o'rtacha javob vaqti | |
| Real-time Latency | 80ms | WebSocket update latency | |
| Concurrent Users | 100 | Bir vaqtda 100 ta foydalanuvchi |