Abstract: With the rapid development of educational technology, the traditional teaching evaluation system in colleges and universities is facing many challenges, especially in processing large-scale educational data and providing real-time feedback. The purpose of this study is to clean and preprocess the original data of multiple sources of information, such as teachers'natural information, teaching process data, supervision scores, student feedback, and course achievement, and use association rule algorithms to complete the screening of strongly correlated indicators, providing new ideas for building a new type of teaching evaluation system in colleges and universities to improve the accuracy and efficiency of the evaluation. This teaching evaluation system can effectively assist teaching management and decision-making, and contribute to improving the quality of teaching.

Key words: Evaluation of university teaching, Association rule algorithm, Data mining, New evaluation system, Model construction, Improvement of teaching quality