A computational modeling strategy was established to predict the tensile strength of polylactic acid (PLA) reinforced with wood particles fabricated using fused deposition modeling (FDM). Three different predictive approaches—Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Response Surface Methodology (RSM)—were applied to estimate the mechanical response of the printed composites. The experimental program was arranged using a Taguchi L27 orthogonal array, considering five influential printing parameters: layer thickness, infill density, printing speed, nozzle temperature, and raster orientation. The analysis revealed that the highest tensile strength of 24.70 MPa was achieved when the printing conditions were set to 95% infill density, 300 μm layer thickness, and a 0° raster angle. ANOVA indicated that raster angle (38.36%) and print speed (32.55%) were the most influential parameters, while infill density and layer thickness had moderate effects. Comparative evaluation of predictive models showed that ANFIS achieved superior accuracy (R² = 0.9906, RMSE = 0.2821, MAPE = 1.06%), outperforming ANN (R² = 0.9376, RMSE = 0.7286, MAPE = 2.67%) and RSM (R² = 0.7318, RMSE = 1.5101, MAPE = 5.39%). Residual and validation analyses confirmed the robustness of ANFIS in capturing nonlinear parameter interactions. The findings indicate that the combined application of various artificial intelligence–driven modeling approaches can greatly minimize reliance on large-scale experimental trials. This integration improves the reliability of prediction results while offering deeper understanding of process parameters, thereby supporting the efficient optimization of manufacturing conditions for durable and environmentally friendly PLA/wood composite materials.
Chauhan et al. (Fri,) studied this question.