Purpose

This study utilizes data analysis supplemented with neural network models to explore the accuracy of prediction models for high school evaluation results in Taiwan and discusses the feasibility of applying data practices in educational evaluations. Considering the distribution of sample sizes for high school evaluations published by the Ministry of Education, this study primarily constructed prediction models for schools receiving grades A and B in the curriculum and teaching dimensions of Taiwan’s high school evaluations. Additionally, it compares the differences in prediction models for general and vocational high schools and identifies the key factors for improving school management.

Design/methodology/approach

The analysis is based on high school evaluation results, linking long-term tracking data from Taiwan’s post-secondary education and school background data from the Ministry of Education’s Statistical Office. The feasibility of data-driven evaluation result prediction was understood using artificial neural network algorithms for model prediction accuracy exploration.

Findings/results

The analysis results reveal that (1) the prediction rate of high school evaluations using database data exceeds 75%, with general high schools at 80% and vocational high schools at 77.9%, indicating the feasibility of using data for school evaluations; (2) the satisfaction level of Taiwanese high school students with their schools is a key factor in predicting the outcomes of school evaluations in curriculum and teaching; (3) the flexibility and diversity of elective courses in high schools are critical predictors of the outcomes of school evaluations in curriculum and teaching; and (4) positive teaching strategies used by teachers in vocational high schools are significant predictors of school curriculum and teaching grades.

Originality/value

This study analyzed data from database information and high school evaluation results by employing neural network methods to establish a predictive model for evaluation outcomes. The research findings indicated the feasibility of using database data to predict evaluation results, validating that using data for school evaluations is a viable direction for development.

Suggestions/implications

This study believes that establishing data databases is valuable because it not only helps the Ministry of Education understand the current educational landscape but also demonstrates the feasibility of developing data-driven school evaluations. This study focused primarily on the dimensions of curriculum and teaching for analysis. If future efforts continue to expand and deepen this approach, it may serve to reduce administrative pressure between educational administrative departments and schools. Simultaneously, the research findings and recommendations offer significant reference values for enhancing the operational efficiency of high school management.