基於學生練習使用回饋之學習成效預測模型與動態題數練習機制;Implement a Dynamic Exercise Mechanism Based on Learning Effectiveness Prediction Model

NCU Institutional Repository > 資訊電機學院 > 網路學習科技研究所 > 博碩士論文 > Item 987654321/81458

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/81458

題名:	基於學生練習使用回饋之學習成效預測模型與動態題數練習機制;Implement a Dynamic Exercise Mechanism Based on Learning Effectiveness Prediction Model
作者:	邱歆雅;Chiou, Shin-Ya
貢獻者:	網路學習科技研究所
關鍵詞:	回饋;動態題數練習機制;精熟學習;適性化學習;Feedback;Mastery Learning;Dynamic Exercise mechanism;Adaptive learning
日期:	2019-08-15
上傳時間:	2019-09-03 15:54:48 (UTC+8)
出版者:	國立中央大學
摘要:	由於許多電腦輔助學習系統會使用回饋來幫助學生學習，但卻發生一個問題：「學生常常濫用回饋」。部分學生不是因為學會所學概念而前往下一階段，而是透過不斷的要求系統回饋而順利進行通關，導致並未學習到該學習的知識內容。因此，本研究希望能透過精熟學習理論與適性化學習的方式，訂定一個精熟門檻，給予每位學生動態的練習題數，讓已經完成基本練習題數並通過精熟門檻的學生，不需要進行額外的練習，可直接進入下一階段，而未通過精熟門檻的學生，給予加強練習，期望能提高學生的精熟程度與對數學概念的理解。　　本研究於「數學島」系統中進行實驗，於實驗第一階段，將學生分為固定練習組（控制組）與動態練習組（實驗組），運用固定練習組學生使用的回饋等級與學習成效評量之歷史資料建立熟練度預測模型。根據預測模型建立動態題數練習機制，動態調整給予動態練習組學生練習的題目數量，希望學生在完成練習後，皆能夠精熟學習內容。而本研究建立不同的預測模型，期望找出能夠準確預測學習成效評量之預測模型。　　研究結果發現，有修正的預測模型比無修正的預測模型準確，使用回饋等級序列與回饋等級總數熟練度預測模型皆能夠準確預測學習成效評量。而就練習總題數來看，動態練習組學生之學習成效評量表現與固定練習組學生接近，由此可知，動態題數練習機制可以精準地預測學習成效評量。 ;Many computer-assisted learning systems adopt immediate feedback to help students learning. However, the problem has been occurred that students using feedback inappropriately. Some students got passed to the next stage by constantly requesting for system feedback instead of actually learning the studying concepts which lead to fewer learn the content knowledge on the mathematics. In this study, we reference by mastery learning theory and adaptive learning to set a threshold of mastery. The threshold of mastery can give each student a dynamic number of exercises. So, students who have completed basic exercises and passed the threshold of mastery can go to the next stage without additional exercises. Otherwise, students who have not passed the threshold of mastery, the system will give them strengthened exercises to improve students′ mastery levels and their understanding of mathematical concepts. This study was conducted in the "Math Island" system. In the first phase of the experiment, students were divided into two groups, the fixed-practice group (control group) and the dynamic-practice group (experimental group). The proficiency prediction model was established by utilizing the historical data of feedback level and learning effectiveness evaluation used by the students in the control group. According to the prediction model, a dynamic exercise mechanism was established and the number of exercises given to the dynamic-exercise group by dynamic adjust. This study hoped that the students will be able to master the learning content after completing the exercises by dynamic exercise mechanism. Moreover, this study also established different prediction models to find an appropriate prediction model that can accurately predict the evaluation of learning effectiveness. The results show that the revised prediction model is more accurate than the non-revised one, and the proficiency prediction model using the sequence of feedback levels and the total number of feedback levels can accurately predict the evaluation of learning effectiveness. As far as the total number of exercises is concerned, the performance of the students in the dynamic-practice group is close to the fixed-practice group. It is found that the dynamic exercise mechanism can accurately predict the evaluation of learning effectiveness.
顯示於類別:	[網路學習科技研究所 ] 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	182	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.....