Research Article
Application of Multiattention Mechanism in Power System Branch Parameter Identification
Table 1
Experimental results of grid dataset on index line susceptance
.
| Line susceptance | None | 50 dB | 30 dB | dropData | dropEdge | 50 dB + DD + DE |
| LinearRegression | 0.0006 | 0.1339 | 0.4402 | 0.1273 | 0.1862 | 0.2569 | SVR | 0.1329 | 0.2161 | 1.4923 | 0.1181 | 0.1780 | 0.2399 | RF | 0.0785 | 0.1558 | 0.2298 | 0.1795 | 0.1898 | 0.2335 | XGBoost | 0.0856 | 0.1732 | 0.2431 | 0.2124 | 0.2373 | 0.2542 | lightGBM | 0.1324 | 0.1988 | 0.3562 | 0.3122 | 0.3242 | 0.3541 | KNN | 0.0943 | 0.1524 | 0.2345 | 0.2462 | 0.2145 | 0.2451 | Bagging | 0.0541 | 0.1451 | 0.2331 | 0.2143 | 0.2364 | 0.2442 |
| FCN | 0.6314 | 0.7603 | 0.8430 | 0.8129 | 0.8724 | 0.8133 | GTN | 0.1413 | 0.1658 | 0.1992 | 0.1696 | 0.1652 | 0.2188 |
|
|
Here we choose RMSE as the comparative index. The highest values for the different metrics are highlighted in bold.
|