Table 1 Classification performance on simulation measurements for event detection, classification and localization.

Power Domain

PMU score⁴³

—

0.266

—

Traditional

1-NN Euclidean^44,45

0.537

0.402

36.465

Machine

1-NN DTW-i^44,45

0.610

0.463

53.928

Learning

1-NN DTW-d^44,45

0.598

0.474

53.709

Methods

MiniRocket⁴²

0.690 ± 0.022

0.208 ± 0.226

53.908 ± 3.358

Vanilla CNN

0.564 ± 0.058

0.168 ± 0.053

40.458 ± 12.686

InceptionTime⁴⁶

0.715 ± 0.040

0.243 ± 0.047

43.743 ± 10.605

MLSTM-FCN⁴⁰

0.742 ± 0.029

0.285 ± 0.023

31.873 ± 5.400

Convolutional

ResNet⁴⁸

0.725 ± 0.049

0.232 ± 0.044

38.578 ± 9.569

Neural

MC-DCNN⁵⁰

0.726 ± 0.019

0.437 ± 0.030

38.107 ± 5.675

Networks

TapNet⁴¹

0.653 ± 0.018

0.397 ± 0.065

58.251 ± 1.974

Fully-connected Neural Network

0.583 ± 0.042

0.245 ± 0.035

54.131 ± 9.964

Other

Vanilla RNN

0.504 ± 0.045

0.224 ± 0.037

57.184 ± 4.285

Deep

LSTM⁵¹

0.544 ± 0.049

0.248 ± 0.043

56.434 ± 2.851

Learning

GRU⁵²

0.653 ± 0.029

0.332 ± 0.062

55.550 ± 2.090

Methods

Vanilla Transformer⁵³

0.612 ± 0.041

0.340 ± 0.090

46.824 ± 0.866