Optimization and Prediction of Tensile Strength in 3D‐Printed PLA/HAp Composites Using Response Surface Methodology and Integrated Machine Learning Technique

ABSTRACT This study aims to optimize the tensile strength of 3D‐printed polylactic acid (PLA) reinforced with hydroxyapatite (HAp). By the integration of response surface methodology (RSM) and machine learning (ML), the most influential process parameters, filler ratio (FR), infill ratio (IR), and layer thickness (LT), were examined for their effects on tensile strength. Experimental confirmation by 3D‐printed specimens established the initial point of the optimal settings. Various ML models, such as least angle regression (LARS), linear regression (LR), ridge regression (RR), and k‐nearest neighbor (KNN), were employed for TS prediction and determination of the optimized level of 3D‐PFs for achieving improved tensile strength. From RSM results, the LARS algorithm founded on the Bagging Ensemble Learning Technique (ELT) can predict and optimize tensile strength with high accuracy. Performance metrics, MAE (0.154), MedAE (0.131), and RMSE (0.164), validated the model's strength. The maximum tensile strength of 46.3 MPa was achieved for optimal parameters for FR of 9%, IR of 30%, and LT of 0.1 mm. LARS had a higher predictive ability with a higher partial dependence value (PDV) for the same parameter values.