本研究主要進行跨平臺資料整合研究，連結教育部高級中等學校學籍資料與中部某教育大學學生學習檔案資料，以類神經網路法分析師培校院選才入學管道、高中學業表現對大學學習成就之預測研究。研究結果顯示，高中學習表現與入學管道能有效預測師資生學習表現。進一步分析發現，有別於以往大學生選才較注重學科表現，本研究透過模型測試指出，入學管道、性別與部分非考科對於預測師資生大學學習表現有其重要效果。最後，本研究結果建構師資生學習表現重要預測因子，提供師培大學作為選才參考。

This research focuses on multi-platform integration, which connects the data of student status in junior high school from the Ministry of Education and student’s learning profile from a university of education in central Taiwan. Neural network is used to analyze the teacher education selection, students’ performance in high school and its projection of students’ performance in university. The results show that students’ performance in high school can be used to predict their performance in university effectively. Moreover, by testing models, it is noticed that in the projection of teacher program students’ performance, some subjects are more important than exam subjects. At last, this research offers the dominant factors of projecting teacher program students’ performance for universities to take into consideration when selecting future teacher program students.