Design and Development of Evaluation of University Curriculum Learning Outcomes Based on Hybrid Deep Learning Architecture

Assessing curriculum learning outcomes is critical for ensuring academic quality and continuous improvement in higher education. This paper presents the design and development of an intelligent evaluation framework for university curriculum learning outcomes using a deep learning model. The system aims to provide accurate, scalable, and automated insights into student performance and learning effectiveness across diverse disciplines. The framework integrates data from course assessments, student feedback, attendance records, online learning behaviors, and final grade distributions. A hybrid deep learning architecture combining Convolutional Neural Networks for structured data analysis and Bidirectional Gated Recurrent Networks networks for textual evaluation was developed. Feature fusion techniques were applied to represent both cognitive outcomes and affective outcomes. A dataset comprising over many course outcome records from multiple departments across three academic years was used to train and validate the model. The proposed system achieved an accuracy of $\mathbf{9 8. 6 \%}$ in predicting learning outcome attainment and strong correlation with expert-evaluated rubrics. The system enables academic institutions to automatically map student learning evidence to program objectives, identify curriculum gaps, and support accreditation and policy-making processes. It offers a transformative approach to data-driven curriculum evaluation, aligned with the goals of Outcome-Based Education and continuous academic improvement