Evaluation of English Teaching Quality in Higher Vocational Colleges Based on Hybrid Neural Networks

Improving the quality of English teaching in higher vocational colleges is essential for enhancing students' global competitiveness and communication skills. This paper proposes the design and development of an intelligent evaluation system for English teaching quality based on Hybrid Neural Networks (HNN). The system integrates pre-trained(PT) model and Long Short-Term Memory (LSTM) models to effectively analyze multidimensional data, including student feedback, classroom performance metrics, teaching content delivery, and engagement indicators. The hybrid architecture leverages PTs feature extraction capabilities and LSTM's sequence modeling strength to capture both static and temporal patterns in educational data. A customized evaluation index system is established, incorporating factors such as teaching effectiveness, language proficiency development, and learning outcomes. Real-world data collected from multiple vocational institutions are used to train and validate the model. Experimental results show that the proposed HNN-based system significantly outperforms traditional evaluation methods in terms of accuracy (accuracy of 99 %), adaptability, and objectivity. This intelligent framework provides actionable insights for academic administrators and educators, contributing to continuous improvement in English education in vocational settings.