Year: 2025
Type: conference-paper
Study design: Clinical Trial
Abstract: Our research investigates how artificial intelligence sentiment analysis technology can be applied in vocational education. We developed a deep learning-based multimodal sentiment analysis model to monitor and analyze learners' emotional states in real time. A 16-week teaching experiment was conducted, involving 300 vocational college students in a quasi-experimental design. The findings revealed three key points. First, our multimodal fusion model demonstrated an accuracy rate of 87.6% in emotion recognition, which was substantially better than single-modal models. Second, students in the experimental group achieved better learning outcomes than those in the control group - their theoretical scores were 7.2 points higher on average (p<0.01), while their practical scores rose by 9.5 points (p<0.001). Third, we observed enhanced learning engagement in the experimental group, with a 35% increase in classroom participation. Their professional identity also showed marked improvement (t=3.45, p<0.01). These findings offer practical guidance and technical solutions for enhancing vocational education quality, contributing to the advancement of educational informatization.
Authors: Xingwang Pi, Yanlun Chen, Qian Liu
Institution: Hunan Automotive Engineering Vocational University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4411744601