Determinants of saturation magnetic flux density in Fe-based metallic glasses: insights from machine-learning models

Fe-based metallic glasses have garnered significant attention due to their low coercivity force and core loss. Enhancing the saturation magnetic flux density (B_s) of Fe-based metallic glasses is crucial for their industry applications. This work constructed a dataset comprising 330 training data and 157 test data. The support vector regression model surpassed the tree-based ensemble models in the test set and demonstrated comparable accuracy to the tree-based ensemble models in the training set. Additionally, we proposed an indicator for B_s based on symbolic regression. This newly proposed indicator exhibits a Pearson correlation coefficient exceeding 0.92 with B_s. The present work provides a simple and accurate formula for predicting the B_s of Fe-based amorphous alloys, demonstrating the effectiveness of machine learning approaches in discovering novel soft magnetic materials.