LAUSR

Abstract In this work we use the random forest model to predict the glass-forming ability (GFA) of metallic alloys by leveraging a previously published study and dataset. The new model with optimized hyperparameters successfully boosts the prediction accuracy by over 12%. The improvement is primarily attributed to the additional critical features (e.g. mixing entropy and total electronegativity) that has been identified. Although the improved model is still far away from being satisfactory (R2 = 0.64) due to the extremely unbalanced nature of the dataset, the new identified features will significantly facilitate the future model development with more and more emerging experimental data.