Protocol to predict mechanical properties of multi-element ceramics using machine learning

Identifying and designing high-performance multi-element ceramics based on trial-and-error approaches are ineffective and expensive. Here, we present a machine-learning-accelerated method for prediction of mechanical properties of multi-element ceramics, based on the density functional theory calculation database. Specific bonding characteristics are used as highly efficient machine learning descriptors. This protocol describes a low-cost, high-efficiency, and reliable workflow for developing advanced ceramics with superior mechanical properties. For complete details on the use and execution of this protocol, please refer to Tang et al. (2021).