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

Table 5

RMSE and MAE based on different inputs.

Data	Training time (s)	Average prediction time (s)
Raw samples	42.86	0.38
RFSs	49.49	0.42
Fusion	59.56	0.46