Scholarlink

<p>The rapid advancement of sodium-ion battery (SIB) technology necessitates efficient methods for evaluating and identifying high-performance cathode materials to accelerate their commercialization. In this context, elucidating the key factors governing rate capability and establishing accurate predictive models are of paramount importance for their rational design. This study addresses this need by integrating machine learning with publicly available data to predict the rate performance of P2-type cathode materials. We employ the maximal information coefficient (MIC) to identify the critical features and construct a high-accuracy prediction model using the random forest (RF) algorithm. To resolve the inherent “black-box” nature of such models, we enhance interpretability through SHapley Additive exPlanations (SHAP) and accumulated local effects (ALE). Finally, 15 representative samples are synthesized to evaluate the accuracy of the model, achieving a low test error of 4.7%. The findings provide theoretical insights and practical guidance for the rapid design, development, and evaluation of high-performance P2-type materials for SIBs.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Rate Performance Prediction of P2-Type Cathode Materials Using a Random Forest Model

  • 期刊论文

  • 发表日期：

    2026年2月10日

查看全文
  • Zijin Cui,
  • Jingyuan Guo,
  • Cheng Wei,
  • Haokun Li,
  • Linzhuang Xing,
  • Hongyu Yang,
  • Yihang Li,
  • Zhimin Li

摘要

The rapid advancement of sodium-ion battery (SIB) technology necessitates efficient methods for evaluating and identifying high-performance cathode materials to accelerate their commercialization. In this context, elucidating the key factors governing rate capability and establishing accurate predictive models are of paramount importance for their rational design. This study addresses this need by integrating machine learning with publicly available data to predict the rate performance of P2-type cathode materials. We employ the maximal information coefficient (MIC) to identify the critical features and construct a high-accuracy prediction model using the random forest (RF) algorithm. To resolve the inherent “black-box” nature of such models, we enhance interpretability through SHapley Additive exPlanations (SHAP) and accumulated local effects (ALE). Finally, 15 representative samples are synthesized to evaluate the accuracy of the model, achieving a low test error of 4.7%. The findings provide theoretical insights and practical guidance for the rapid design, development, and evaluation of high-performance P2-type materials for SIBs.