Key Factors and Predictive Models of Digital Collaborative Education Based on Machine Learning.

Area of Science:

• Educational Technology
• Machine Learning in Education
• Teacher Professional Development

Background:

• Digital collaborative education is crucial for improving teaching quality and fostering talent development through family-school-community partnerships.
• Key drivers and predictive models for digital collaborative education effectiveness are not well understood.
• Teachers' digital literacy is a critical, yet underexplored, factor in digital collaborative education.

Purpose of the Study:

• To identify key factors influencing digital collaborative education among primary and secondary school teachers.
• To develop predictive models for digital collaborative education effectiveness using machine learning.
• To analyze the heterogeneity of these factors across different teacher demographics.

Main Methods:

• Employed machine learning algorithms, including gradient boosting regression trees (GBRT) and random forest, to identify influential factors and build predictive models.
• Utilized SHapley Additive exPlanations (SHAP) for comprehensive explanatory analysis and accumulated local effects (ALE) plots for single-feature interpretation.
• Focused on primary and secondary school teachers as the research subjects.

Main Results:

• Random forest model demonstrated superior performance in predicting digital collaborative education effectiveness compared to other models.
• Digital application, digital academic assessment, digital instructional implementation, and digital teaching design were identified as the most significant predictors.
• Digital application emerged as the strongest predictor, followed by professional development, with nonlinear relationships observed across various demographic groups.

Conclusions:

• The study provides empirical evidence supporting the advancement of digital collaborative education.
• Digital literacy, particularly digital application and professional development, significantly impacts digital collaborative education outcomes.
• Findings offer valuable insights for targeted teacher professional development programs to enhance digital collaborative education.