Evaluation of the Effectiveness of Artificial Intelligence-Assisted Nuclear Electronics Teaching Based on Multiple Regression
With the accelerated empowerment of higher education by artificial intelligence (AI) technology, how to effectively assess its impact on the teaching effectiveness of specific courses has become a topic worthy of attention. This study takes the core course of nuclear technology discipline at a certain university, “Nuclear Electronics”, as a case, and adopts a quasi-experimental design to compare the academic performance of 100 students in the AI-integrated group and the non-AI-integrated group. Based on the multivariate linear regression model, while controlling for pretest scores (atomic physics scores), continuous assessment scores, and self-efficacy, it systematically assesses the impact of AI-assisted teaching on students' academic achievements in the course, and simultaneously identifies the key predictive factors affecting academic performance. The results show that AI-assisted teaching has a significant positive net effect on students' academic achievements $(\mathrm{B}=4.68, \mathrm{p}=0.008)$, among which continuous assessment scores are the strongest predictive factor $(\mathrm{B}=1.25, \mathrm{p}<0.001)$. This study provides a quantifiable and scalable empirical paradigm for AI empowerment in engineering education curriculum reform, and has reference value for promoting the intelligent teaching transformation of nuclear technology-related and similar engineering majors.