<![CDATA[In higher education, especially science and biology, digital technology in project-based learning (PjBL) environments has improved student engagement and learning outcomes. technological, AI, and lecturer assistance have been studied in PjBL, but few have used Artificial Neural Networks (ANN) to analyze the complicated interactions between technological acceptance variables and student engagement. ANN is used to predict students' attitudes toward technology (ATT), intention to use technology (INT), actual use of technology in PjBL (AU-PjBL), and student engagement (SE) based on PEOU, PU, and Lecturer Support. Biology education students at Universitas Jambi completed a 35-item Likert-scale questionnaire. We created four ANN models: Model A (PU, PEOU → ATT), Model B (PU, ATT → INT), Model C (INT, LS → AU-PjBL), and Model D (AU-PjBL, LS → Each model was trained and tested using ten network configurations. Model performance was assessed using Root Mean Square Error (RMSE), and input variable relevance was determined via sensitivity analysis. All ANN models have low RMSE values for training and testing datasets, indicating good predicting accuracy. According to sensitivity analysis, PU predicts ATT better than PEOU, ATT predicts INT better than PU, INT predicts AU-PjBL better than LS, and AU-PjBL predicts SE better than LS. These data emphasize that students' perceived utility, positive attitudes, intention, and technology use drive biology PjBL involvement. The paper highlights ANN as a powerful analytical tool for modeling non-linear and interdependent relationships in technology-enhanced PjBL and gives practical implications for developing meaningful technology use and engagement learning environments. Keywords: Artifical neural network; actual use of technology; lecturer support; project-based learning; biology education.]]>
Copyrights © 2025
