Fig. 4: K-fold cross-validation method for the SOH estimation of each cell in G-I Dataset.
From: Efficient estimating and clustering lithium-ion batteries with a deep-learning approach
a The estimated capacity versus cycle number of cell G-I#6 when the model uses base features at Vu of 2.60 V, 3.05 V and 3.20 V, respectively. The density distribution of the capacity estimation error is shown as inserted figure. b Model prediction accuracy with the input of base feature at Vu of 3.05 V. c The average MAE and RMSE for the model with combined features at an upper voltage limit (Vu) of 3.05 V.
