Development of the Teaching Quality Evaluation Model for Higher Vocational Colleges Based on Optuna-LightGBM

In the context of big data, an abundance of data has emerged in the field of educational assessment, offering substantial potential value for educators. This paper introduces the LightGBM algorithm into the teaching quality evaluation model, utilizes the Optuna automatic hyperparameter optimization framework for parameter tuning, and compares its performance with that of decision tree, GBDT, Adaboost, random forest and XGBoost. The evaluation indicators including accuracy, precision, recall rate, F1 value, AUC and ROC curve are employed to objectively differentiate teaching quality in a scientific and professional manner. The results demonstrate that the Optuna-LightGBM based teaching quality evaluation model exhibits superior predictive accuracy, generalization ability and operational efficiency compared to other models. This approach provides a scientific and objective means for distinguishing between good and poor teaching quality while offering instructive suggestions for improving teaching work.