Research on teaching effect prediction model and system implementation based on time series knowledge graph and association rule mining

With the rapid development of educational informatization, accurately predicting teaching results is of great significance in optimizing teaching strategies and improving teaching quality. Traditional teaching effectiveness evaluation methods primarily rely on static data, making it challenging to capture the dynamic characteristics of time series and potential correlations. This study integrates a time series knowledge graph and time series association rule mining technology to construct a multidimensional data modeling framework for the teaching process. By integrating time series data, such as students’ learning behavior logs, curriculum resource interaction records, and periodic evaluation results, a knowledge graph is constructed that includes knowledge point association, learning path evolution, and achievement fluctuation laws. Based on the Apriori algorithm, a time series association rule mining model suitable for teaching scenarios is developed to identify typical strong association rules. The experimental results show that the prediction accuracy of this model for teaching effectiveness is 89.7%, which is 13.2% points higher than that of the traditional logistic regression model, and the recall rate is 85.3%. It can effectively identify the key factors affecting teaching effectiveness and provide a scientific basis for personalized teaching intervention.