A novel approach to predict the effect of chemical composition and thermo-mechanical processing parameters on Cu-Ni-Si alloys using a hybrid deep learning and ensemble learning model

The study presents a novel hybrid deep learning and ensemble learning (DL-EL) model to predict the effects of chemical composition and thermo-mechanical processing on the properties of Cu-Ni-Si alloys. The model integrates various input parameters like chemical composition and thermo-mechanical processing parameters and aims to predict key output properties such as mechanical properties and electrical conductivity. This study addresses gaps in existing research by providing a comprehensive model for Cu-Ni-Si alloys and integrating thermo-mechanical processing parameters with chemical composition in the predictive model. The research demonstrates the model's superior predictive performance, with near-perfect R2 values on both training and test sets. The hybrid DL-EL model was compared with three other machine learning models and its efficacy was assessed using R2 value, offering new insights into the performance of these alloys. A feature importance analysis was conducted to identify the most influential features of the model. This work contributes to the material science field by providing insights into optimizing alloy composition and processing, leveraging machine learning to enhance material design. (c) 2017 Elsevier Inc. All rights reserved.