Abstract
Lithium-metal batteries (LMBs) are emerging as a promising next-generation energy storage due to their exceptionally high energy density. However, accurately predicting their performance remains challenging because of the complex degradation mechanisms. In this study, a machine learning (ML) framework is proposed that combines electrochemical impedance spectroscopy (EIS) with the XGBoost algorithm to develop two predictive models: one for estimating capacity degradation and another for detecting the knee point (KP)-a critical inflection point in the degradation trajectory. SHapley Additive exPlanations (SHAP) analysis is employed to interpret feature importance, revealing that low-frequency imaginary impedance components-associated with diffusion-limited processes such as lithium depletion and accumulation-are most influential for capacity estimation. Conversely, high-frequency features related to charge transfer resistance play a dominant role in the KP detection. To reduce data complexity and improve model efficiency, the input by selecting specific frequency points based on SHAP values is further optimized. The optimized models exhibit comparable or improved accuracy compared to those using the whole EIS data and have reasonable performance on unseen test data. The findings highlight that EIS-based ML models can accurately forecast heaslth of LMBs, providing deeper insights into their aging processes and enhancing battery management strategies.