Design and Empirical Study of a Process-oriented Evaluation Model for the Integration of Vocational Education, Innovation and Entrepreneurship in Higher Vocational Education Based on LSTM-Attention
Year: 2025
Type: conference-paper
Abstract: Aiming at the three major evaluation pain points in the integration of vocational education, innovation and entrepreneurship (VIE integration) teaching in higher vocational education—"result-oriented, delayed feedback, and lack of temporal modeling"—this study introduces the LSTM-Attention architecture into the higher vocational education scenario for the first time. It constructs a process-oriented evaluation system with "five dimensions and twelve indicators" and designs a three-level dynamic feedback loop of "weekly data collection - monthly curriculum adjustment - semester weight update". The study completed model training and empirical research based on 12-week multi-source data of 320 students: ①The model achieved an F1 score of 0.94 for early warning in weak dimensions, a mean absolute error (MAE) of only 2.31 for score prediction, and the Attention mechanism contributed a 4.8% improvement; ② The application of the system increased teachers' intervention response speed by 62% and students' award-winning rate in the "Internet +" Competition by27%; ③The score of interdisciplinary collaboration increased by an average of 12.7%, showing a strong correlation with competition results (r=0.76). The results indicate that this model realizes the accurate connection between "weekly" dynamic evaluation and teaching improvement, and provides an interpretable and promotable technical solution for the process-oriented evaluation of VIE integration in higher vocational education.
Authors: Zhongkun Wu, Nayan Li
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W7126218271
