Optimization of teacher evaluation indicator system based on fuzzy-DEMATEL-BP

Area of Science:

• Educational Measurement and Statistics
• Artificial Intelligence in Education
• Data Science

Background:

• Teacher evaluation indicator systems are crucial for educational quality.
• Existing systems often suffer from complexity and fuzzy uncertainties.
• Simplifying and optimizing indicators is essential for accurate and efficient evaluation.

Purpose of the Study:

• To propose a novel method for optimizing teacher evaluation indicator systems.
• To address complex causal relationships and fuzzy uncertainties in indicator selection.
• To establish a precise and rational teacher evaluation index system.

Main Methods:

• Combined Triangular Fuzzy Decision-making Trial and Evaluation Laboratory Model (Fuzzy-DEMATEL) with Backpropagation Neural Network (BP).
• Utilized DEMATEL to analyze causal relationships and Fuzzy logic for uncertainty handling.
• Employed BP neural network for objective data training to mitigate subjective errors.

Main Results:

• The proposed method effectively simplifies complex assessment systems and identifies key indicators.
• Empirical analysis on TIMSS 2019 data showed a reduction in indicators by 28%-30%.
• Achieved superior accuracy and reduced indicator quantity compared to Multi-Criteria Decision Making (MCDM).

Conclusions:

• The Fuzzy-DEMATEL and BP neural network combination offers a robust approach to optimizing teacher evaluation indicators.
• This method enhances the interpretability and precision of teacher evaluation systems.
• The optimized indicator system leads to more accurate and efficient teacher assessments.