Design and Development of Higher Education Quality Evaluation based on Data Mining Technology

With the sudden invasion of the Internet into the lives of all, online assessment of teaching has become a critical tool in gauging the quality of institutional instruction. Data mining methods are utilized to extract valuable information from these assessments, providing a solid scientific basis for improving teaching quality and administrative control. This paper suggests an intelligent framework for data mining encompassing both subjective and objective assessments. Over 1,75,900 test data of evaluation were gathered from different educational institutions which is employed for training and model testing. The raw data are pre-processed, annotated, and cleaned first. Then, for subjective test data, a model integrating Text-to-Text Transfer Transformer(T5) Model and pre-trained models like Gated Recurrent Neural Networks (GRNN) networks is built to make sentiment tendencies predictions with an accuracy of 94.76% and proving the effectiveness of the model. For objective evaluation data, the Apriori algorithm is utilized to extract association rules, from which significant rules are chosen for study. This study efficiently investigates teaching evaluation data, offering technical help for improving higher teaching quality and formulating educational reform strategies.