Research on the Multimodal Digital Teaching Quality Data Evaluation Model Based on Fuzzy BP Neural Network

Area of Science:

• Artificial Intelligence
• Educational Technology
• Machine Learning

Background:

• Multimodal digital teaching requires robust quality evaluation methods.
• Traditional BP neural networks face challenges with initial weights and subjectivity.
• Existing evaluation models may lack accuracy and efficiency.

Purpose of the Study:

• To develop and evaluate an optimized fuzzy BP neural network for multimodal digital teaching quality assessment.
• To improve the prediction accuracy and convergence speed of neural network models in educational evaluation.
• To reduce the subjectivity inherent in neural network learning for teaching quality data.

Main Methods:

• Proposed a fuzzy BP neural network model enhanced with an adaptive variation genetic algorithm (GA-BP).
• Utilized the entropy value method to obtain objective a priori guidance samples.
• Optimized initial weights and thresholds of the BP neural network using the genetic algorithm.
• Compared GA-BP model performance against original BP, GA, and BSA algorithms.

Main Results:

• The GA-BP neural network evaluation model demonstrated higher accuracy in multimodal digital teaching quality assessment.
• Optimization using the adaptive variation genetic algorithm reduced training time and improved convergence.
• The entropy method successfully guided the BP neural network, reducing sample learning subjectivity.
• Comparative analysis confirmed the superior performance of the GA-BP model over other tested methods.

Conclusions:

• The GA-BP neural network presents a more accurate and feasible solution for multimodal digital teaching quality evaluation.
• Genetic algorithm optimization effectively addresses BP neural network limitations, enhancing predictive capabilities.
• The integration of entropy-based guidance improves the objectivity and reliability of the evaluation model.