應用資料探勘技術探究我國高中生適性學習影響因素

APPLICATION OF DATA MINING THEORY TO INVESTIGATE FACTORS IMPACTING HIGH SCHOOL STUDENTS' ADAPTIVE LEARNING IN TAIWAN

研究目的：本文主要探究我國高級中等學校教學環境、教師活化教學、學生家庭背景與學習歷程等相關因素對於學生適性學習的影響，其結果可提供教育政策研擬與教育資源投入之參據。根據上述，本研究目的如下：（1）應用不同資料探勘理論，分析我國高中學生適性學習之關鍵影響因素；（2）探究我國高中學生適性學習相關因素之影響排序；（3）研究結果提供中央教育主管機關，作為教育政策擬定之參據。研究設計／方法／取徑：本研究採問卷調查法，調查研究對象為105學年度高一、二學生，採問卷調查法進行。調查設計以學校為母體，採多階段分層隨機抽樣。本次調查共計抽出121所學校，每校共發出60份問卷，高一與高二各30份問卷，回收有效樣本數為6,628人。本研究將所蒐集之樣本進行資料探勘理論—類神經網路、k－平均演算分析及決策樹分析，探究不同因素對於適性學習的影響效果。研究發現或結論：1.學校教學環境是影響學生適性學習的主要因素。2.學生家庭期望對於學生適性學習效果影響程度較低。3.學生學習歷程與教師活化教學對於學生適性學習的影響尚待後續評估。研究原創性／價值：本研究係以評估高中學生適性學習的前提下，運用資料探勘方法—類神經網路、k－平均演算分析及決策樹。從學校環境、教師教學、學生學習歷程、家庭期望等，分析影響高中學生適性學習的關鍵因素，解決過往線性模式分析對於因素影響程度的不確定性。教育政策建議或實務意涵：本研究分析結果顯示，學校教學環境是學生適性學習關鍵影響因素。然，檢視我國近年來高中均優質化政策之推展，加上完全免試入學政策的推動，著實改善目前高中學校教學環境，對於協助學生適性學習有若干幫助。然下一階段有關學校教學環境提升，應配合特色課程發展為主，協助各校建構其特色與專長，讓學生在高中學習階段具備更多元選擇，達成我國發展適性學習的教育目的。 Purpose: The purpose of this study is to investigate the teaching environment, teacher creative teaching, students' family background and learning process on students' adaptive learning. The results of this study can be used to provide suggestions for education policies and the input of educational resources. The purpose is (1) applying different machine learning theory to investigate the key factors impacting high school students' adaptive learning (2) investigate the orders of factors impacting high school students' adaptive learning (3) results can be used to provide suggestions for policy makers for educational authorities. Design/methodology/approach: This study uses the questionnaire survey to investigate the 10th and 11th graders in 2016. The study use the multilevel random sampling design, a total of 121 schools is sampled and each school receives 60 questionnaires. Each grade has 30 questionnaires, totally 6,628 students finish the questionnaires. Data mining approach, including neural network, k-means, and decision tree analysis are used to investigate the factors impacting adaptive learning. Findings: 1. School teaching environment is the main factor, which impacts students' adaptive learning. 2. Students' family expectation has low impact on students' adaptive learning. 3. The impact of student learning process and teachers' teaching on students' adaptive learning needs further investigation. Originality/value: This study applies the machine learning approach neural network, k-means, decision tree analysis to investigate the factors impacting students' adaptive learning, including school environment, teacher teaching, students' learning process and family background. The uncertainty level for previous linear model analysis can be solved. Implications for policy/practice: The result of this study shows that school environment is the key factor impacting students' adaptive learning. Current completely non-exam entrance policy can assist students' adaptive learning; however, for promoting the school teaching environment in the next level, schools should develop their own feature curriculum. Students can have multiple options for deciding their high school, so the purposes of adaptive learning can be achieved.