Research Article
Spatiotemporal Approaches for Quality Control and Error Correction of Atmospheric Data through Machine Learning
Table 5
Comparison of MLQC-NT results on raw data and interpolated data using SVR.
| Weather element | Raw data | Interpolated data | Correlation coefficient | RMSE | 3σ error | Time (sec) | Correlation coefficient | RMSE | 3σ error | Time (sec) | #err/#total | Rateerr(%) | #err/#total | Rateerr(%) |
| Temperature | 1.0000 | 0.0586 | 177/10050 | 1.76 | 1363 | 1.0000 | 0.0553 | 192/9665 | 1.99 | 1487 | Humidity | 0.9995 | 0.6424 | 504/10050 | 5.01 | 6881 | 0.9997 | 0.4690 | 185/6429 | 2.88 | 3298 | UV-rays | 0.9919 | 1.7307 | 4039/10050 | 40.19 | 5995 | 0.9919 | 1.7307 | 4039/10050 | 40.19 | 5995 | PM2.5 | 0.9211 | 8.1466 | 382/10050 | 3.80 | 5400 | 0.9834 | 3.6097 | 372/10013 | 3.72 | 3044 | Solar radiation | 0.9388 | 0.0039 | 52/10050 | 0.52 | 3605 | 0.9388 | 0.0039 | 52/10050 | 0.52 | 3605 | u | 0.7592 | 0.5801 | 560/10050 | 5.57 | 1414 | 0.7772 | 0.3563 | 386/7254 | 5.32 | 543 | | 0.7650 | 0.7378 | 591/10050 | 5.88 | 1496 | 0.7848 | 0.4244 | 451/7254 | 6.22 | 580 |
|
|