Online Learning & Drift Detection
প্রতি ৩ ঘণ্টায় স্বয়ংক্রিয় model retrain · উন্নত হলেই deploy
Retraining Strategy
Frequency
৩ ঘণ্টা
Auto scheduled
Drift Threshold
১০%
CPU বা RAM mean drift
Deploy Policy
উন্নত হলে
Only if RMSE improves
মোট Retrain
50
এখন পর্যন্ত
🚀 সফল Deploy
15
মোট 50 retrain এর মধ্যে
📉 সর্বশেষ CPU RMSE
5.29%
আগে ছিল: 5.12%
⚡ সর্বশেষ CPU Drift
23.8%
Threshold: 10% · V1011
সর্বশেষ Retrain এর প্রভাব — Real-time Database থেকে
🧠 CPU Forecast — V1011 · 2026-10-12 20:19
আগে
5.12%
RMSE
→
পরে
5.29%
RMSE
3% ↑
খারাপ
Drift: 23.8%
· R²: 0.832 → 0.820
· Samples: 45,125
💾 RAM Forecast — V1011 · 2026-10-12 20:19
আগে
1.68%
RMSE
→
পরে
1.67%
RMSE
0% ↓
উন্নত
RAM Drift: 23.4%
· Deploy: ✅ হয়েছে
RMSE উন্নতির গ্রাফ
CPU RMSE (%)
RAM RMSE (%)
Retrain ইতিহাস (সর্বশেষ 50টি)
| সময় | VPS | CPU আগে | CPU পরে | RAM আগে | RAM পরে | CPU Drift | Samples | Status |
|---|---|---|---|---|---|---|---|---|
| 2026-10-12 20:19 | V1011 | 5.1203% | 5.2923% ↑ | 1.6770% | 1.6743% ↓ | 23.8% | 45,125 | ✅ Deploy |
| 2026-10-12 17:18 | V1011 | 5.1206% | 5.1632% ↑ | 1.6771% | 1.6816% ↑ | 14.3% | 45,089 | ⏭️ Skip |
| 2026-10-12 14:18 | V1011 | 5.1148% | 5.1844% ↑ | 1.6800% | 1.6773% ↓ | 47.4% | 45,053 | ✅ Deploy |
| 2026-10-12 11:17 | V1011 | 5.1106% | 5.2218% ↑ | 1.6746% | 1.6858% ↑ | 43.5% | 45,017 | ⏭️ Skip |
| 2026-10-12 08:17 | V1011 | 5.1108% | 5.1634% ↑ | 1.6752% | 1.6846% ↑ | 55.8% | 44,981 | ⏭️ Skip |
| 2026-10-12 05:16 | V1011 | 5.1081% | 5.1810% ↑ | 1.6581% | 1.6685% ↑ | 61.6% | 44,945 | ⏭️ Skip |
| 2026-10-12 02:16 | V1011 | 5.1087% | 5.2138% ↑ | 1.6642% | 1.6587% ↓ | 51.6% | 44,909 | ✅ Deploy |
| 2026-10-11 23:15 | V1011 | 5.1079% | 5.1866% ↑ | 1.6650% | 1.6716% ↑ | 49.1% | 44,873 | ⏭️ Skip |
| 2026-10-11 20:15 | V1011 | 5.1052% | 5.2040% ↑ | 1.6604% | 1.6704% ↑ | 44.6% | 44,837 | ⏭️ Skip |
| 2026-10-11 17:14 | V1011 | 5.1044% | 5.2203% ↑ | 1.7622% | 1.6608% ↓ | 32.4% | 44,800 | ✅ Deploy |
| 2026-10-11 14:14 | V1011 | 5.1035% | 5.1797% ↑ | 1.7692% | 1.7864% ↑ | 34.3% | 44,764 | ⏭️ Skip |
| 2026-10-11 11:13 | V1011 | 5.1114% | 5.1094% ↓ | 1.7684% | 1.7817% ↑ | 28.4% | 44,728 | ✅ Deploy |
| 2026-10-11 08:13 | V1011 | 5.1101% | 5.2488% ↑ | 1.7690% | 1.7808% ↑ | 35.8% | 44,692 | ⏭️ Skip |
| 2026-10-11 05:11 | V1011 | 5.1127% | 5.1296% ↑ | 1.7632% | 1.7718% ↑ | 51.6% | 44,656 | ⏭️ Skip |
| 2026-10-11 02:11 | V1011 | 5.1147% | 5.2043% ↑ | 1.7634% | 1.7790% ↑ | 31.8% | 44,620 | ⏭️ Skip |
| 2026-10-10 23:10 | V1011 | 5.1187% | 5.1994% ↑ | 1.7640% | 1.7831% ↑ | 10.3% | 44,584 | ⏭️ Skip |
| 2026-10-10 20:10 | V1011 | 5.1106% | 5.2223% ↑ | 1.7654% | 1.7643% ↓ | 35.5% | 44,547 | ✅ Deploy |
| 2026-10-10 17:09 | V1011 | 5.1089% | 5.1752% ↑ | 1.7662% | 1.7727% ↑ | 42.3% | 44,511 | ⏭️ Skip |
| 2026-10-10 14:09 | V1011 | 5.1638% | 5.1144% ↓ | 1.7684% | 1.7709% ↑ | 57.9% | 44,475 | ✅ Deploy |
| 2026-10-10 11:08 | V1011 | 5.1636% | 5.1628% ↓ | 1.8149% | 1.7595% ↓ | 57.9% | 44,439 | ✅ Deploy |
| 2026-10-10 08:07 | V1011 | 5.2374% | 5.1623% ↓ | 1.8185% | 1.8219% ↑ | 62.6% | 44,403 | ✅ Deploy |
| 2026-10-10 05:06 | V1011 | 5.2340% | 5.2516% ↑ | 1.8022% | 1.8084% ↑ | 70.9% | 44,367 | ⏭️ Skip |
| 2026-10-10 02:06 | V1011 | 5.2347% | 5.3085% ↑ | 1.8026% | 1.8086% ↑ | 59.1% | 44,331 | ⏭️ Skip |
| 2026-10-09 23:05 | V1011 | 5.2334% | 5.2691% ↑ | 1.8039% | 1.8074% ↑ | 28.5% | 44,295 | ⏭️ Skip |
| 2026-10-09 20:04 | V1011 | 5.2373% | 5.3454% ↑ | 1.8044% | 1.8070% ↑ | 30.6% | 44,258 | ⏭️ Skip |
| 2026-10-09 17:04 | V1011 | 5.2393% | 5.2871% ↑ | 1.8051% | 1.8065% ↑ | 38.3% | 44,222 | ⏭️ Skip |
| 2026-10-09 14:03 | V1011 | 5.2477% | 5.2369% ↓ | 1.8057% | 1.8079% ↑ | 29.9% | 44,186 | ✅ Deploy |
| 2026-10-09 11:02 | V1011 | 5.2502% | 5.3278% ↑ | 1.7935% | 1.8076% ↑ | 32.6% | 44,150 | ⏭️ Skip |
| 2026-10-09 08:02 | V1011 | 5.2542% | 5.3107% ↑ | 1.7929% | 1.7981% ↑ | 37.6% | 44,114 | ⏭️ Skip |
| 2026-10-09 05:01 | V1011 | 5.2565% | 5.3387% ↑ | 1.7718% | 1.7710% ↓ | 39.2% | 44,078 | ✅ Deploy |
| 2026-10-09 02:00 | V1011 | 5.2588% | 5.3694% ↑ | 1.7732% | 1.7940% ↑ | 25.1% | 44,042 | ⏭️ Skip |
| 2026-10-08 23:00 | V1011 | 5.2589% | 5.3164% ↑ | 1.7739% | 1.7818% ↑ | 27.9% | 44,005 | ⏭️ Skip |
| 2026-10-08 19:59 | V1011 | 5.2720% | 5.2579% ↓ | 1.7745% | 1.7867% ↑ | 35.9% | 43,969 | ✅ Deploy |
| 2026-10-08 16:58 | V1011 | 5.2698% | 5.2729% ↑ | 1.7752% | 1.7780% ↑ | 40.2% | 43,933 | ⏭️ Skip |
| 2026-10-08 13:58 | V1011 | 5.2729% | 5.2866% ↑ | 1.7762% | 1.7872% ↑ | 39.5% | 43,897 | ⏭️ Skip |
| 2026-10-08 10:57 | V1011 | 5.2834% | 5.2766% ↓ | 1.7658% | 1.7714% ↑ | 32.0% | 43,861 | ✅ Deploy |
| 2026-10-08 07:56 | V1011 | 5.2874% | 5.3064% ↑ | 1.7662% | 1.7801% ↑ | 44.1% | 43,825 | ⏭️ Skip |
| 2026-10-08 04:55 | V1011 | 5.2930% | 5.4209% ↑ | 1.7604% | 1.7665% ↑ | 55.5% | 43,789 | ⏭️ Skip |
| 2026-10-08 01:54 | V1011 | 5.2917% | 5.3245% ↑ | 1.7651% | 1.7752% ↑ | 46.6% | 43,752 | ⏭️ Skip |
| 2026-10-07 22:53 | V1011 | 5.2994% | 5.2895% ↓ | 1.7737% | 1.7640% ↓ | 46.8% | 43,716 | ✅ Deploy |
| 2026-10-07 19:53 | V1011 | 5.2977% | 5.3642% ↑ | 1.7777% | 1.7928% ↑ | 49.0% | 43,680 | ⏭️ Skip |
| 2026-10-07 16:52 | V1011 | 5.2975% | 5.2933% ↓ | 1.7782% | 1.7899% ↑ | 50.8% | 43,644 | ✅ Deploy |
| 2026-10-07 13:51 | V1011 | 5.2971% | 5.3435% ↑ | 1.7787% | 1.7875% ↑ | 46.2% | 43,608 | ⏭️ Skip |
| 2026-10-07 10:50 | V1011 | 5.2999% | 5.3648% ↑ | 1.7727% | 1.7842% ↑ | 49.2% | 43,572 | ⏭️ Skip |
| 2026-10-07 07:49 | V1011 | 5.2991% | 5.3750% ↑ | 1.7770% | 1.7890% ↑ | 63.2% | 43,535 | ⏭️ Skip |
| 2026-10-07 04:49 | V1011 | 5.3004% | 5.3480% ↑ | 1.7517% | 1.7620% ↑ | 65.1% | 43,499 | ⏭️ Skip |
| 2026-10-07 01:49 | V1011 | 5.3036% | 5.3446% ↑ | 1.7432% | 1.7545% ↑ | 41.2% | 43,463 | ⏭️ Skip |
| 2026-10-06 22:48 | V1011 | 5.2782% | 5.3561% ↑ | 1.6446% | 1.6526% ↑ | 33.6% | 43,427 | ⏭️ Skip |
| 2026-10-06 19:47 | V1011 | 5.2743% | 5.3369% ↑ | 1.6445% | 1.6565% ↑ | 17.5% | 43,391 | ⏭️ Skip |
| 2026-10-06 16:46 | V1011 | 5.2805% | 5.3250% ↑ | 1.6454% | 1.6540% ↑ | 17.3% | 43,355 | ⏭️ Skip |
কীভাবে কাজ করে
📊
ডেটা সংগ্রহ
প্রতি ৩০ সেকেন্ডে নতুন metrics
📈
Drift শনাক্ত
নতুন vs পুরনো mean (threshold 10%)
🔄
Retrain
XGBoost নতুন ডেটা দিয়ে train
✅
যাচাই
পুরনো vs নতুন RMSE তুলনা
🚀
Deploy
উন্নত হলেই replace