Prediction of tensile strength in AA7075/SiC composites using random forest and support vector machine models

In the present study, regression approaches of machine learning were implemented to predict the tensile strength of AA7075/SiC composites fabricated under different process parameters. SiC composition, compaction pressure, sintering temperature, and sintering time were considered as independent variables, while tensile strength was considered as a dependent response. This study presents an exploratory data analysis through parametric distribution using pairplots and boxplots, showing how process parameters affect the tensile strength of AMCs. GridSearchCV was used for hyperparameter tuning in the following two optimized supervised regression models: Random Forest and Support Vector Machine. The coefficient of determination (R²), mean absolute error (MAE), and root mean square error (RMSE) were used to evaluate model performance. The Random Forest model demonstrated better predictive performance than the SVM model, as evident from R² = 0.965, MAE = 7.81, and RMSE = 9.50 versus R² = 0.838, MAE = 16.51, and RMSE = 20.43, respectively. This implies that the Random Forest algorithm is robust enough to provide a better generalization on nonlinear relationships between processing parameters and the tensile strength of composites and thus is reliable for optimization of composite manufacturing processes.