Answers are grounded in this paper and cite their sources.
Three issues in assessing higher education administration quality are data uncertainty, non-linear relationships among indicators, and lack of interpretability. Traditional linear weighting and shallow sequential hybrid methods cannot effectively solve these problems. This paper proposes an evaluation model integrating an interval type-2 fuzzy layer with convolutional neural network, bidirectional long short-term memory, and attention mechanism. The fuzzy layer captures high-order uncertainty via differentiable membership functions, and the outputs are merged with original features for end-to-end training. On an anonymised archival dataset of 4782 records from a single institution, the model achieves accuracy of 94.12%, F1-score of 93.97%, and area under the curve of 0.973, surpassing the best baseline fuzzy deep learning method by 2.89 and 4.07 percentage points. Ablation studies indicate that removing any module, including the fuzzy layer, degrades performance on the present dataset. Attention visualisation and fuzzy rules provide exploratory, model-level interpretability for administrators; their value as validated management guidance remains untested. This work offers a precise and intuitive evaluation tool for higher education management.
No takes yet. Share an insight, caveat, or question.
Jinnan Yin (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: