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
6
মোট 50 retrain এর মধ্যে
📉 সর্বশেষ CPU RMSE
3.68%
আগে ছিল: 3.68%
⚡ সর্বশেষ CPU Drift
17.2%
Threshold: 10% · V1009
সর্বশেষ Retrain এর প্রভাব — Real-time Database থেকে
🧠 CPU Forecast — V1009 · 2026-10-12 20:18
আগে
3.68%
RMSE
→
পরে
3.68%
RMSE
0% ↑
খারাপ
Drift: 17.2%
· R²: 0.459 → 0.458
· Samples: 44,987
💾 RAM Forecast — V1009 · 2026-10-12 20:18
আগে
2.92%
RMSE
→
পরে
3.01%
RMSE
3% ↑
খারাপ
RAM Drift: 63.5%
· Deploy: ⏭️ Skip
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 20:18 | V1009 | 3.6816% | 3.6848% ↑ | 2.9175% | 3.0120% ↑ | 17.2% | 44,987 | ⏭️ Skip |
| 2026-10-12 17:18 | V1009 | 3.6796% | 3.6818% ↑ | 2.8967% | 3.0516% ↑ | 23.2% | 44,951 | ⏭️ Skip |
| 2026-10-12 14:18 | V1009 | 3.6802% | 3.6835% ↑ | 2.8870% | 2.9949% ↑ | 18.8% | 44,915 | ⏭️ Skip |
| 2026-10-12 11:17 | V1009 | 3.6770% | 3.6806% ↑ | 2.8704% | 2.9596% ↑ | 25.2% | 44,879 | ⏭️ Skip |
| 2026-10-12 08:17 | V1009 | 3.6775% | 3.6812% ↑ | 2.8539% | 3.0207% ↑ | 18.9% | 44,843 | ⏭️ Skip |
| 2026-10-12 05:16 | V1009 | 3.6732% | 3.6758% ↑ | 2.8364% | 2.9555% ↑ | 33.6% | 44,807 | ⏭️ Skip |
| 2026-10-12 02:15 | V1009 | 3.6836% | 3.6867% ↑ | 2.8230% | 2.9457% ↑ | 27.7% | 44,771 | ⏭️ Skip |
| 2026-10-11 23:15 | V1009 | 3.6841% | 3.6873% ↑ | 2.8086% | 2.9502% ↑ | 6.2% | 44,735 | ⏭️ Skip |
| 2026-10-11 20:14 | V1009 | 3.6567% | 3.6619% ↑ | 2.7901% | 2.8450% ↑ | 13.6% | 44,698 | ⏭️ Skip |
| 2026-10-11 17:14 | V1009 | 3.6545% | 3.6579% ↑ | 2.7769% | 2.8700% ↑ | 14.4% | 44,662 | ⏭️ Skip |
| 2026-10-11 14:13 | V1009 | 3.6491% | 3.6510% ↑ | 2.7636% | 2.8680% ↑ | 4.0% | 44,626 | ⏭️ Skip |
| 2026-10-11 11:13 | V1009 | 3.6356% | 3.6360% ↑ | 2.7489% | 2.8239% ↑ | 8.1% | 44,590 | ⏭️ Skip |
| 2026-10-11 08:12 | V1009 | 3.6310% | 3.6309% ↓ | 2.7173% | 2.7618% ↑ | 28.2% | 44,554 | ✅ Deploy |
| 2026-10-11 05:11 | V1009 | 3.6322% | 3.6320% ↓ | 2.6980% | 2.8083% ↑ | 13.0% | 44,518 | ✅ Deploy |
| 2026-10-11 02:10 | V1009 | 3.6351% | 3.6368% ↑ | 2.6872% | 2.8709% ↑ | 21.4% | 44,482 | ⏭️ Skip |
| 2026-10-10 23:10 | V1009 | 3.6393% | 3.6406% ↑ | 2.6703% | 2.8594% ↑ | 8.6% | 44,445 | ⏭️ Skip |
| 2026-10-10 20:09 | V1009 | 3.6233% | 3.6273% ↑ | 2.6568% | 2.7345% ↑ | 11.3% | 44,409 | ⏭️ Skip |
| 2026-10-10 17:09 | V1009 | 3.5930% | 3.5971% ↑ | 2.6380% | 2.7255% ↑ | 6.1% | 44,373 | ⏭️ Skip |
| 2026-10-10 14:08 | V1009 | 3.5709% | 3.5738% ↑ | 2.6252% | 2.7843% ↑ | 7.1% | 44,337 | ⏭️ Skip |
| 2026-10-10 11:07 | V1009 | 3.5635% | 3.5658% ↑ | 2.6087% | 2.7193% ↑ | 23.4% | 44,301 | ⏭️ Skip |
| 2026-10-10 08:07 | V1009 | 3.5723% | 3.5755% ↑ | 2.5876% | 2.6853% ↑ | 29.3% | 44,265 | ⏭️ Skip |
| 2026-10-10 05:06 | V1009 | 3.5690% | 3.5743% ↑ | 2.5785% | 2.6712% ↑ | 23.8% | 44,229 | ⏭️ Skip |
| 2026-10-10 02:05 | V1009 | 3.5656% | 3.5673% ↑ | 2.5694% | 2.6512% ↑ | 29.9% | 44,193 | ⏭️ Skip |
| 2026-10-09 23:04 | V1009 | 3.5698% | 3.5721% ↑ | 2.5472% | 2.6353% ↑ | 9.6% | 44,156 | ⏭️ Skip |
| 2026-10-09 20:04 | V1009 | 3.5527% | 3.5550% ↑ | 2.5205% | 2.5738% ↑ | 21.5% | 44,120 | ⏭️ Skip |
| 2026-10-09 17:03 | V1009 | 3.5538% | 3.5595% ↑ | 2.5003% | 2.6305% ↑ | 2.2% | 44,084 | ⏭️ Skip |
| 2026-10-09 14:02 | V1009 | 3.5322% | 3.5358% ↑ | 2.4774% | 2.5718% ↑ | 12.3% | 44,048 | ⏭️ Skip |
| 2026-10-09 11:02 | V1009 | 3.5272% | 3.5281% ↑ | 2.4557% | 2.5524% ↑ | 22.7% | 44,012 | ⏭️ Skip |
| 2026-10-09 08:01 | V1009 | 3.4804% | 3.4786% ↓ | 2.3959% | 2.4767% ↑ | 27.0% | 43,976 | ✅ Deploy |
| 2026-10-09 05:01 | V1009 | 3.4840% | 3.4875% ↑ | 2.3671% | 2.4719% ↑ | 27.3% | 43,940 | ⏭️ Skip |
| 2026-10-09 02:00 | V1009 | 3.4833% | 3.4857% ↑ | 2.3554% | 2.4533% ↑ | 29.7% | 43,904 | ⏭️ Skip |
| 2026-10-08 22:59 | V1009 | 3.4839% | 3.4865% ↑ | 2.3222% | 2.4203% ↑ | 3.4% | 43,867 | ⏭️ Skip |
| 2026-10-08 19:58 | V1009 | 3.4980% | 3.4995% ↑ | 2.2839% | 2.4298% ↑ | 48.6% | 43,831 | ⏭️ Skip |
| 2026-10-08 16:58 | V1009 | 3.4586% | 3.4619% ↑ | 2.2418% | 2.3476% ↑ | 23.1% | 43,795 | ⏭️ Skip |
| 2026-10-08 13:57 | V1009 | 3.4659% | 3.4668% ↑ | 2.2012% | 2.2965% ↑ | 6.1% | 43,759 | ⏭️ Skip |
| 2026-10-08 10:56 | V1009 | 3.4382% | 3.4392% ↑ | 2.1632% | 2.2020% ↑ | 13.4% | 43,723 | ⏭️ Skip |
| 2026-10-08 07:55 | V1009 | 3.4307% | 3.4311% ↑ | 2.1427% | 2.1913% ↑ | 25.6% | 43,687 | ⏭️ Skip |
| 2026-10-08 04:54 | V1009 | 3.4340% | 3.4332% ↓ | 2.1124% | 2.1584% ↑ | 28.9% | 43,650 | ✅ Deploy |
| 2026-10-08 01:53 | V1009 | 3.4429% | 3.4434% ↑ | 2.0917% | 2.1615% ↑ | 19.9% | 43,614 | ⏭️ Skip |
| 2026-10-07 22:53 | V1009 | 3.4595% | 3.4631% ↑ | 2.0463% | 2.1161% ↑ | 4.4% | 43,578 | ⏭️ Skip |
| 2026-10-07 19:52 | V1009 | 3.4608% | 3.4640% ↑ | 2.0274% | 2.1302% ↑ | 22.2% | 43,542 | ⏭️ Skip |
| 2026-10-07 16:51 | V1009 | 3.4614% | 3.4612% ↓ | 2.0190% | 2.0958% ↑ | 12.8% | 43,506 | ✅ Deploy |
| 2026-10-07 13:50 | V1009 | 3.4334% | 3.4318% ↓ | 2.0047% | 2.1114% ↑ | 11.1% | 43,470 | ✅ Deploy |
| 2026-10-07 10:50 | V1009 | 3.4262% | 3.4267% ↑ | 1.9733% | 2.2940% ↑ | 21.2% | 43,433 | ⏭️ Skip |
| 2026-10-07 07:49 | V1009 | 3.4315% | 3.4324% ↑ | 1.9283% | 2.2595% ↑ | 7.2% | 43,397 | ⏭️ Skip |
| 2026-10-07 04:49 | V1009 | 3.3974% | 3.4016% ↑ | 1.8739% | 2.1143% ↑ | 8.2% | 43,361 | ⏭️ Skip |
| 2026-10-07 01:48 | V1009 | 3.3779% | 3.3797% ↑ | 1.8200% | 2.0516% ↑ | 20.7% | 43,325 | ⏭️ Skip |
| 2026-10-06 22:47 | V1009 | 3.3770% | 3.3775% ↑ | 1.7728% | 2.0194% ↑ | 7.3% | 43,289 | ⏭️ Skip |
| 2026-10-06 19:46 | V1009 | 3.3679% | 3.3702% ↑ | 1.7409% | 1.9910% ↑ | 20.9% | 43,253 | ⏭️ Skip |
| 2026-10-06 16:45 | V1009 | 3.3669% | 3.3690% ↑ | 1.7234% | 1.9903% ↑ | 12.3% | 43,217 | ⏭️ 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