Evaluation of teaching quality in probability theory and mathematical statistics based on an improved BP neural network algorithm

This study proposes an enhanced Backpropagation (BP) neural network algorithm to evaluate the teaching quality of probability theory and mathematical statistics courses, leveraging the principles of smart education. Traditional evaluation methods often suffer from subjectivity and limited dimensionality, which can hinder the accuracy and effectiveness of assessments. To address these challenges, we optimize the BP neural network by introducing adaptive learning rates and momentum factors, thus improving convergence speed and prediction accuracy. By integrating diverse input indicators, including student feedback, examination results, and expert assessments, the model constructs a comprehensive and data-driven evaluation system. The experimental results demonstrate that the improved algorithm not only achieves higher precision and stability compared to conventional methods but also effectively reduces overfitting. This approach aligns with the goals of smart education by facilitating data-driven decision-making and supporting continuous improvement in educational quality monitoring and pedagogical optimization.