Crystal System Classification of Phosphate Cathode Materials Using Machine Learning for Advanced Lithium‐Ion Battery

ABSTRACT We employed machine learning classification algorithms to predict the crystal systems, specifically triclinic, monoclinic, and orthorhombic, associated with Li–P–TM–O; (TM = Mn, Fe Co, Ni, V) based‐phosphate cathodes. Feature evaluation revealed that cathode properties depend on the crystal structure, and optimized classification strategies yield better predictability. Gradient Boosting Machines, Extremely Randomized Trees, and Random Forest‐based ensemble machine learning algorithms have demonstrated the best predictive capabilities for crystal systems in the Monte Carlo cross‐validation test. Additionally, sequential forward selection (SFS) is performed to find the utmost critical features influencing the accuracy of prediction for different machine learning models, with Volume, Band gap, and Sites as input features. The ensemble machine learning algorithms with 80.69% for Random Forest, 78.96% for Extremely Randomized Tree, and 80.40% for Gradient Boosting Machine approaches lead to the maximum accuracy toward crystallographic classification with stability, and the predicted materials can be the possible cathodes for lithium ion batteries. More importantly, the present approach can be extended to other groups of materials.