State-of-Health Estimation for Lithium-Ion Battery Based on an Attention-Based CNN-GRU Model with Reconstructed Feature Series

Table 4

RMSE and MAE based on different inputs.

Metrics	Data	cell5	cell6	cell7	cell8	Average
RMSE (%)	Raw samples	0.876	0.946	0.859	0.830	0.878
	RFSs	0.776	0.865	0.753	0.732	0.782
	Fusion	0.529	0.630	0.629	0.541	0.582
MAE (%)	Raw samples	0.753	0.880	0.704	0.730	0.767
	RFSs	0.665	0.735	0.67	0.608	0.669
	Fusion	0.446	0.590	0.586	0.475	0.524