An artificial intelligence method for comprehensive evaluation of preschool education quality

Area of Science:

• Education
• Artificial Intelligence
• Educational Technology

Background:

• Preschool education quality evaluation faces qualitative challenges.
• Existing fuzzy comprehensive evaluation and neural network models have limitations.

Purpose of the Study:

• To propose a quantitative method for comprehensive quality evaluation in preschool education.
• To develop an improved Feedforward Neural Network (FNN) model for automated evaluation.

Main Methods:

• Established a set of indicators for kindergarten student comprehensive quality evaluation.
• Utilized analytic hierarchy process for index weighting.
• Developed a Feedforward Neural Network (FNN) model integrating fuzzy logic and neural network characteristics.
• Improved FNN convergence speed by using similarity measure to enhance hidden layer nodes.

Main Results:

• The proposed FNN model enhances the learning ability and expressive power of fuzzy evaluations.
• The improved FNN model demonstrates faster convergence compared to traditional methods.
• Effectiveness and feasibility of the enhanced FNN for automated comprehensive quality evaluation were verified.

Conclusions:

• The improved FNN model offers a viable solution for automated and accurate comprehensive quality evaluation in preschool education.
• This approach addresses limitations of previous methods, paving the way for more objective educational assessments.