結合早期預警和影片推薦機制改善學生學習成效;Combining early warning and video recommendation mechanisms to improve students’ learning performance

NCU Institutional Repository > 資訊電機學院 > 資訊工程研究所 > 博碩士論文 > Item 987654321/81065

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/81065

題名:	結合早期預警和影片推薦機制改善學生學習成效;Combining early warning and video recommendation mechanisms to improve students’ learning performance
作者:	翁健軒;Weng, Jian-Xuan
貢獻者:	資訊工程學系
關鍵詞:	適性化推薦;主成分迴歸;評分策略;學習成效預測;多元分類;學習風險識別;Adaptive Recommendation;PCR;Grading Policy;Students’ Learning Performance Prediction;Multi-class Classification;At-risk Students Identification
日期:	2019-07-02
上傳時間:	2019-09-03 15:32:30 (UTC+8)
出版者:	國立中央大學
摘要:	本論文以ilearning線上學習平台的一門系統程式課程為研究目標，實驗包含兩大主軸：早期預警及影片教材推薦。研究中提出一套推薦機制，結合早期預警之結果和影片推薦清單，在適當的時機給予學生，以提升學生的學習效率，並改善學生的學習成效。早期預警包含干預時機的選擇及預測準確度兩大要素，本論文蒐集近五年學生的線上學習歷程、課程大綱、作業及考試成績，從中萃取出特徵，利用主成分迴歸(Principal Component Regression, PCR)建立學習成效預測模型，再觀察pMSE的預測指標，找出最佳的干預時間點。另一方面，本研究比較八種分類演算法，以accuracy、precision、recall、f1-score和AUC作為評估模型好壞的指標，並區分出高風險學生，最後於推薦清單中，給予適性化的警示話語提醒學生。影片教材推薦是透過學生的線上學習歷程及測驗作答情形，給予學生適性化推薦清單，以期望達到改善學生學習成效的目標。實驗結果顯示，搭配問卷分群結果發現，在中度學習動機和中度學習興趣的學生族群當中，在進步成績的表現上，實驗組學生顯著優於控制組，代表本推薦機制對於特定族群的學生具有顯著的影響。 ;This study is based on a System Programming course of the ilearning online learning platform. The experiment consists of two main purposes: early warning and video recommendation. In this study, a recommendation mechanism was proposed, which combined the results of early warning and the list of recommended videos to give students at the right time to improve students′ learning efficiency and improve their learning outcomes. Early warning includes two factors: the timing of intervention and the accuracy of prediction. This paper collects online learning history, syllabus, homework and test scores of students in the past five years, extracts features from it, and uses Principal Component Regression (PCR) to establish a learning effectiveness prediction model, and then observe the prediction indicators named pMSE to find the best intervention time point. On the other hand, this study compares eight classification algorithms, using accuracy, precision, recall, f1-score, and AUC as indicators to evaluate the quality of the model, and distinguishes high-risk students, and finally gives appropriateness alert words to remind students. The video recommendation is to give students a suitable recommendation list through the online learning process and test answering situation of the students, in order to achieve the goal of improving the learning performance of the students. The experimental results show that the results of the grouping with the questionnaire found that among the student populations with moderate learning motivation and moderate learning interest, the experimental group students were significantly better than the control group. It means that the recommendation mechanism for students of specific groups of students has a significant impact.
顯示於類別:	[資訊工程研究所] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	75	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....