Intelligent Systems with Applications (Nov 2023)
Exploring the application of teaching evaluation models incorporating association rules and weighted naive Bayesian algorithms
Abstract
Teaching evaluation is a judgment of the value of teachers’ teaching and students’ learning, and has become an important part of teaching management and teaching processes in universities. However, the workflow of implementing teaching evaluation is relatively cumbersome, often requiring the completion of a large amount of data calculation tasks. Therefore, how to apply modern science and technology to establish a comprehensive, objective and feasible teaching evaluation system and optimize the evaluation process is an important issue that urgently needs to be solved. The study first uses the Apriori algorithm to explore the correlation between evaluation indicators and results, and then optimizes the teaching evaluation indicators. On this basis, incremental learning is used to improve the classification training ability of the weighted naive Bayesian algorithm, and it is combined with the Apriori algorithm for teaching evaluation. The results show that the fused algorithm takes only 50 seconds to process 500 transactions, and the running speed improves rapidly. As the minimum support threshold decreases, the increase in the time required by the algorithm gradually decreases, resulting in a higher running speed. In the self-built university teaching evaluation database, compared with the BP (Error Back Propagation) algorithm, the combined algorithm has a relatively small fluctuation of accuracy in classifying teaching data, stable at 80 % to 95 %. Meanwhile, in the efficiency comparison, the algorithm requires a slow increase in time as the amount of data increases, resulting in higher efficiency. The application of the algorithm after the fusion of the two in teaching evaluation has not only high accuracy, but also higher efficiency, providing reference technical support for the optimization of university teaching evaluation.
