透過您的圖書館登入
IP:3.16.66.206
  • 學位論文

在數位學習環境下K-sets非監督式分群演算法之應用

Application of K-sets Unsupervised Clustering in Digital Learning Environment

指導教授 : 白宏達
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


近年來,眼動儀(eye tracker)普遍地運用在關於學習的研究上,有許多研究結果顯示,人們於學習中, 關於眼動資料的特徵與最後的學習成效有密切的關係,而本研究就是透過眼動儀來擷取相關的特徵, 最後透過聚類分析(Cluster analysis)來分類出兩個聚類(Cluster)將其視為學習成效好的以及不好的兩 個分類,從中找出,關於眼動資料相對於學習成效最具代表的特徵。 本文使用聚類分析(Cluster analysis)去處理眼動儀分析蒐集之數據,其中的特徵如平均凝視(Gaze) 時間、題目回視次數、凝視最大深度、以及受試時所花的時間等等,由許多的受試者形成資料庫,而 聚類分析(Cluster analysis)則採用最簡單的 K-means 演算法,以及 K-sets 演算法,透過比較兩者演算 法,來完成我們的研究。

並列摘要


In recent years, eye trackers have been commonly used in learning research. There have been many studies showing that the features of eye-movement are strongly related to learning performance. In this study, we use an eye tracker to retrieve related features and cluster analysis to classify learning performance into two clusters comprising respectively of good and bad results, in order to discover the most representative features. Here cluster analysis is applied to the collected eye tracker data including the average fixation time, regression numbers, fixation maximum depth, time spent on the subject. We compare the K-means and K-sets algorithms to complete our study.

參考文獻


[1] C.-S. Chang, W.-J. Liao and Y.-S. Chen, “A mathematical theory for
Network Science and Engineering.
[2] J. A. Hartigan and M. A. Wong, “A k-means clustering algorithm,”
Journal of the Royal Statistical Society. Series C (Applied Statistics),
[3] K. Rayner, “Eye movements in reading and information processing:

延伸閱讀