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
9
মোট 50 retrain এর মধ্যে
📉 সর্বশেষ CPU RMSE
6.13%
আগে ছিল: 6.13%
⚡ সর্বশেষ CPU Drift
22.0%
Threshold: 10% · V1001
সর্বশেষ Retrain এর প্রভাব — Real-time Database থেকে
🧠 CPU Forecast — V1001 · 2026-10-12 20:19
আগে
6.13%
RMSE
→
পরে
6.13%
RMSE
0% ↑
খারাপ
Drift: 22.0%
· R²: 0.602 → 0.601
· Samples: 44,567
💾 RAM Forecast — V1001 · 2026-10-12 20:19
আগে
2.80%
RMSE
→
পরে
2.85%
RMSE
2% ↑
খারাপ
RAM Drift: 1.0%
· 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 20:19 | V1001 | 6.1260% | 6.1334% ↑ | 2.8021% | 2.8453% ↑ | 22.0% | 44,567 | ⏭️ Skip |
| 2026-10-12 17:18 | V1001 | 6.1287% | 6.1349% ↑ | 2.8005% | 2.8407% ↑ | 6.0% | 44,531 | ⏭️ Skip |
| 2026-10-12 14:18 | V1001 | 6.1260% | 6.1308% ↑ | 2.8014% | 2.8430% ↑ | 33.5% | 44,495 | ⏭️ Skip |
| 2026-10-12 11:17 | V1001 | 6.1290% | 6.1312% ↑ | 2.7992% | 2.8407% ↑ | 43.5% | 44,459 | ⏭️ Skip |
| 2026-10-12 08:17 | V1001 | 6.1434% | 6.1482% ↑ | 2.8004% | 2.8417% ↑ | 38.7% | 44,423 | ⏭️ Skip |
| 2026-10-12 05:16 | V1001 | 6.1493% | 6.1582% ↑ | 2.8039% | 2.8452% ↑ | 44.5% | 44,387 | ⏭️ Skip |
| 2026-10-12 02:16 | V1001 | 6.1538% | 6.1626% ↑ | 2.8131% | 2.8529% ↑ | 26.7% | 44,351 | ⏭️ Skip |
| 2026-10-11 23:15 | V1001 | 6.1670% | 6.1670% ↑ | 2.8112% | 2.8112% ↑ | 1.6% | 44,315 | ⏭️ Skip |
| 2026-10-11 20:15 | V1001 | 6.1670% | 6.1670% ↑ | 2.8112% | 2.8112% ↑ | 6.9% | 44,279 | ⏭️ Skip |
| 2026-10-11 17:14 | V1001 | 6.1670% | 6.1729% ↑ | 2.8112% | 2.8492% ↑ | 4.4% | 44,242 | ⏭️ Skip |
| 2026-10-11 14:14 | V1001 | 6.1681% | 6.1726% ↑ | 2.8155% | 2.8550% ↑ | 7.7% | 44,206 | ⏭️ Skip |
| 2026-10-11 11:13 | V1001 | 6.1663% | 6.1715% ↑ | 2.8201% | 2.8618% ↑ | 18.5% | 44,170 | ⏭️ Skip |
| 2026-10-11 08:13 | V1001 | 6.1628% | 6.1653% ↑ | 2.8277% | 2.8669% ↑ | 23.2% | 44,134 | ⏭️ Skip |
| 2026-10-11 05:11 | V1001 | 6.1612% | 6.1625% ↑ | 2.8183% | 2.8608% ↑ | 14.2% | 44,098 | ⏭️ Skip |
| 2026-10-11 02:10 | V1001 | 6.1645% | 6.1679% ↑ | 2.8114% | 2.8571% ↑ | 19.0% | 44,062 | ⏭️ Skip |
| 2026-10-10 23:10 | V1001 | 6.1666% | 6.1738% ↑ | 2.8155% | 2.8543% ↑ | 34.0% | 44,026 | ⏭️ Skip |
| 2026-10-10 20:09 | V1001 | 6.1729% | 6.1780% ↑ | 2.8201% | 2.8660% ↑ | 34.7% | 43,989 | ⏭️ Skip |
| 2026-10-10 17:09 | V1001 | 6.1665% | 6.1737% ↑ | 2.8390% | 2.8800% ↑ | 30.8% | 43,953 | ⏭️ Skip |
| 2026-10-10 14:08 | V1001 | 6.1715% | 6.1794% ↑ | 2.8958% | 2.9007% ↑ | 31.6% | 43,917 | ⏭️ Skip |
| 2026-10-10 11:07 | V1001 | 6.1666% | 6.1685% ↑ | 2.9034% | 2.8980% ↓ | 33.9% | 43,881 | ✅ Deploy |
| 2026-10-10 08:07 | V1001 | 6.1691% | 6.1758% ↑ | 2.9283% | 2.9332% ↑ | 37.7% | 43,845 | ⏭️ Skip |
| 2026-10-10 05:06 | V1001 | 6.1658% | 6.1706% ↑ | 2.9339% | 2.9376% ↑ | 42.2% | 43,809 | ⏭️ Skip |
| 2026-10-10 02:05 | V1001 | 6.1729% | 6.1790% ↑ | 2.9360% | 2.9406% ↑ | 26.7% | 43,773 | ⏭️ Skip |
| 2026-10-09 23:05 | V1001 | 6.1753% | 6.1793% ↑ | 2.9411% | 2.9478% ↑ | 31.4% | 43,736 | ⏭️ Skip |
| 2026-10-09 20:04 | V1001 | 6.1842% | 6.1879% ↑ | 2.9426% | 2.9485% ↑ | 37.2% | 43,700 | ⏭️ Skip |
| 2026-10-09 17:04 | V1001 | 6.1918% | 6.2025% ↑ | 2.9408% | 2.9411% ↑ | 26.2% | 43,664 | ⏭️ Skip |
| 2026-10-09 14:03 | V1001 | 6.1900% | 6.1929% ↑ | 2.9387% | 2.9383% ↓ | 29.3% | 43,628 | ✅ Deploy |
| 2026-10-09 11:02 | V1001 | 6.1898% | 6.1942% ↑ | 2.9355% | 2.9396% ↑ | 19.9% | 43,592 | ⏭️ Skip |
| 2026-10-09 08:02 | V1001 | 6.1769% | 6.1819% ↑ | 2.9383% | 2.9417% ↑ | 41.1% | 43,556 | ⏭️ Skip |
| 2026-10-09 05:01 | V1001 | 6.1803% | 6.1834% ↑ | 2.9458% | 2.9520% ↑ | 25.7% | 43,520 | ⏭️ Skip |
| 2026-10-09 02:00 | V1001 | 6.1868% | 6.1929% ↑ | 2.9554% | 2.9487% ↓ | 30.8% | 43,484 | ✅ Deploy |
| 2026-10-08 22:59 | V1001 | 6.2014% | 6.2061% ↑ | 2.9806% | 2.9717% ↓ | 1.0% | 43,447 | ✅ Deploy |
| 2026-10-08 19:58 | V1001 | 6.1933% | 6.1936% ↑ | 2.9838% | 2.9842% ↑ | 17.3% | 43,411 | ⏭️ Skip |
| 2026-10-08 16:58 | V1001 | 6.1989% | 6.2017% ↑ | 2.9855% | 2.9960% ↑ | 8.4% | 43,375 | ⏭️ Skip |
| 2026-10-08 13:57 | V1001 | 6.1834% | 6.1916% ↑ | 2.9832% | 2.9914% ↑ | 27.6% | 43,339 | ⏭️ Skip |
| 2026-10-08 10:56 | V1001 | 6.1796% | 6.1890% ↑ | 2.9868% | 2.9914% ↑ | 10.4% | 43,303 | ⏭️ Skip |
| 2026-10-08 07:56 | V1001 | 6.1894% | 6.1935% ↑ | 2.9765% | 2.9835% ↑ | 23.6% | 43,267 | ⏭️ Skip |
| 2026-10-08 04:55 | V1001 | 6.1890% | 6.1947% ↑ | 2.9714% | 2.9673% ↓ | 21.1% | 43,231 | ✅ Deploy |
| 2026-10-08 01:54 | V1001 | 6.1762% | 6.1762% ↑ | 2.9718% | 2.9718% ↑ | 5.4% | 43,194 | ⏭️ Skip |
| 2026-10-07 22:53 | V1001 | 6.1762% | 6.1837% ↑ | 2.9718% | 2.9738% ↑ | 28.2% | 43,158 | ⏭️ Skip |
| 2026-10-07 19:52 | V1001 | 6.1789% | 6.1789% ↑ | 2.9824% | 2.9824% ↑ | 5.3% | 43,122 | ⏭️ Skip |
| 2026-10-07 16:52 | V1001 | 6.1789% | 6.1911% ↑ | 2.9824% | 2.9843% ↑ | 0.7% | 43,086 | ⏭️ Skip |
| 2026-10-07 13:51 | V1001 | 6.1701% | 6.1851% ↑ | 2.9884% | 2.9960% ↑ | 21.8% | 43,050 | ⏭️ Skip |
| 2026-10-07 10:50 | V1001 | 6.1725% | 6.1833% ↑ | 2.9820% | 2.9804% ↓ | 39.5% | 43,014 | ✅ Deploy |
| 2026-10-07 07:49 | V1001 | 6.1745% | 6.1879% ↑ | 2.9711% | 2.9761% ↑ | 34.0% | 42,977 | ⏭️ Skip |
| 2026-10-07 04:49 | V1001 | 6.1682% | 6.1804% ↑ | 2.9783% | 2.9837% ↑ | 26.9% | 42,941 | ⏭️ Skip |
| 2026-10-07 01:48 | V1001 | 6.1745% | 6.1842% ↑ | 2.9979% | 2.9946% ↓ | 13.7% | 42,905 | ✅ Deploy |
| 2026-10-06 22:47 | V1001 | 6.1698% | 6.1814% ↑ | 3.0048% | 3.0047% ↓ | 34.0% | 42,869 | ✅ Deploy |
| 2026-10-06 19:47 | V1001 | 6.1716% | 6.1871% ↑ | 3.0092% | 3.0182% ↑ | 21.2% | 42,833 | ⏭️ Skip |
| 2026-10-06 16:46 | V1001 | 6.1727% | 6.1866% ↑ | 3.0266% | 3.0115% ↓ | 2.8% | 42,797 | ✅ Deploy |
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