Construction and Application of a Data Mining-Based Evaluation System for Higher Vocational Colleges
Year: 2026
Type: conference-paper
Abstract: To enhance the scientific rigor and adaptability of quality evaluation in higher vocational colleges, this study constructs a multi-level comprehensive evaluation system based on data mining. Centered on four dimensions—institutional teaching investment, teaching process, learning behavior, and educational outcomes—it systematically designs an evaluation framework encompassing data preprocessing, unsupervised clustering, supervised modeling, and intelligent fusion. The study utilizes multi-source heterogeneous data from 12 higher vocational colleges spanning 2020–2023. By integrating feature normalization, feature subspace screening, and hierarchical structural modeling, it achieves precise classification and comprehensive scoring of evaluation subjects. The clustering results achieved a contour coefficient of 0.612. In supervised learning, the gradient-boosted tree model attained the highest R² value (0.842) and lowest MSE (4.12), validating the model's effectiveness in feature modeling and nonlinear fitting. The final output of dimensional scores and comprehensive results demonstrated strong discriminative power and stability, providing a data foundation and methodological support for constructing a dynamic quality monitoring system for vocational education.
Institution: Binzhou Polytechnic
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W7155642690