AI-Empowered Blended Precision Teaching Model for Computer Courses: Construction and Empirical Study

Abstract

Aiming at the prominent problems in computer courses at application-oriented undergraduate and vocational colleges-namely, significant student divergence, abstract theoretical content, delayed error correction in practice, and one‐dimensional evaluation-this study, against the backdrop of educational digital transformation, integrates artificial intelligence techniques including learner analytics, intelligent code assessment, and personalised resource recommendation. It designs and implements a "three‐dimensional, four‐stage" AI‐driven precision blended teaching model. A one‐semester comparative teaching experiment was conducted between two parallel classes: the experimental class adopted the new model while the control class followed the traditional approach. Effects were examined through quantitative analysis of final grades, questionnaire surveys, and independent‐samples t‐tests. Results show that the model significantly improves students' course performance (P < 0.05), with the excellence rate rising by 17.32 percentage points and the failure rate decreasing by 12.68 percentage points. Students' autonomous learning, programming practice, and classroom engagement are also markedly enhanced. This study provides a replicable and quantitatively grounded reference for the smart teaching reform of computer‐related courses.