Table 2 Mean and standard deviation of residuals for different forecasting models in Case 1 and Case 2.

From: Residual LSTM-based short duration forecasting of polarization current for effective assessment of transformers insulation

Model

(a) Case 1

(b) Case 2

Mean residual

Std dev residual

Mean residual

Std dev residual

Attention LSTM

\(-1.6700 \times 10^{-11}\)

\(5.1807 \times 10^{-11}\)

\(-1.4836 \times 10^{-11}\)

\(4.8007 \times 10^{-11}\)

LSTM

\(-2.9950 \times 10^{-11}\)

\(7.0636 \times 10^{-11}\)

\(-2.5608 \times 10^{-11}\)

\(6.8006 \times 10^{-11}\)

GRU

\(-1.3892 \times 10^{-11}\)

\(4.5987 \times 10^{-11}\)

\(-1.1619 \times 10^{-11}\)

\(4.2006 \times 10^{-11}\)

CNN

\(-1.1850 \times 10^{-11}\)

\(3.2181 \times 10^{-11}\)

\(-1.1475 \times 10^{-11}\)

\(3.7358 \times 10^{-11}\)

Residual LSTM

\(1.2017 \times 10^{-11}\)

\(2.7876 \times 10^{-11}\)

\(1.1260 \times 10^{-11}\)

\(2.3006 \times 10^{-11}\)

Back to article page