探索基於預測產生教材的個人化干預對學生學習成效和SRL學習策略的影響;Exploring the effect of prediction-based content intervention on students’ learning performance and SRL learning strategy.

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/89743

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/89743

Title:	探索基於預測產生教材的個人化干預對學生學習成效和SRL學習策略的影響;Exploring the effect of prediction-based content intervention on students’ learning performance and SRL learning strategy.
Authors:	張介韋;Chang, Jei-Wei
Contributors:	資訊工程學系
Keywords:	精準教育;學習分析;個人化干預;自然語言處理;機器學習;Precision Education;Learning Analytics;Personalized Intervention;Natural Language Processing;Machine Learning
Date:	2022-07-11
Issue Date:	2022-10-04 11:58:15 (UTC+8)
Publisher:	國立中央大學
Abstract:	在過去的研究中發現透過行為分析相較於考試成績可以更好的預測學生的學習成效。但是透過學習分析找出高風險的學生後該如何為他們提供及時的干預以提高學生的學習成效是重要的。本研究根據大學生在線上學習環境的課程中的行為日誌，透過機器學習的方法識別學生的概念是否精熟。另外，進一步研究了從學習日誌中提取的哪些特徵會影響學生各概念的精熟程度，分類器識別學生為不精熟之後，透過學生的學習日誌來找出學生缺乏哪些行為、進而根據學生的情況提供不同個人化的教材當作干預。在個人化教材的生成上，透過自然語言處理的方式幫助教師生成個人化教材以減輕教師的負擔。研究結果顯示分類器識別了大多數有風險的學生，也就是說根據學生的線上學習行為可以找出有風險的學生，進一步幫助教師可以了解學生的學習情況。在不同的概念上影響學生概念精熟與否的特徵也不相同，且實施個人化干預對於提高學生的學習成效有顯著的幫助，在干預措施之成效上，實驗組的學生在程式概念上相較於干預前更為精熟，且相較於控制組在程式測驗成績上的表現顯著較佳。;In past studies, behavioral analysis has been found to be a better predictor of student learning outcomes than test scores. However, after identifying at-risk students through learning analysis, it is important to provide them with timely interventions to improve students′ learning outcomes. This study uses machine learning to identify students′ conceptual mastery level based on the learning logs of college students in courses that support an online learning environment. In addition, we further studied which features extracted from the learning logs would affect the mastery level of each concept of. After the classifier identified the students as not mastery, the students′ learning logs were used to find out which behaviors the students lacked, and then based on it provide personalized materials as interventions. In the personalized teaching materials generation, we help teachers generate personalized teaching materials through natural language processing to reduce the burden on teachers. The results of the study show that the classifier identified most of at-risk students, which means that at-risk students can be identified based on students′ online learning log, which further helps teachers understand students′ learning situations. Different concepts have different characteristics that affect students′ concept mastery level, and the implementation of individualized interventions can significantly help improve students′ learning effectiveness. In terms of the effectiveness of intervention measures, the students in the experimental group are similar in terms of program concepts. More proficient than before the intervention, and significantly better than the control group in programming test.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	49	View/Open

社群 sharing

Loading...