以學習風格為基礎之學習資源推薦機制研究

The Study of Learning Resource Recommandation Mechanism Based on Learning Style

管怡婷(Yi-Ting Kuan)

指導教授 ： 李維平

中原大學/商學院/資訊管理研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400470

摘要

摘 要
相較於傳統教學，網路教學模式有著不受時間空間的限制、降低教育成本、提高教育品質等優點。目前網路教學之個人化相關研究通常僅加上資訊技術提供個人化之服務，而忽略了學習者的學習風格也是影響學習的重要因素。所以本研究欲加入學習風格關聯推薦機制，推薦更符合學習者喜好的學習資源。
本研究首先利用Kolb的學習風格問卷分析學習者的學習風格(Learning Style)，將學習者分成四種學習風格，然後請學習者上網站學習。再將不同學習風格學習者的瀏覽記錄做關聯規則分析，找出四種學習風格學習者的瀏覽行為模式。最後學習者瀏覽記錄與規則比對後，推薦未閱讀文章及相關文章。
並以實驗來驗證學習風格關聯推薦機制效果，學習資源以全民英檢中級閱讀能力測驗為範圍，並將文章分成5大主題，共有50篇文章。實驗對象為中原大學資管系一、二年級共117人，分成3組。A組為無推薦，B組為關聯推薦，C組為有學習風格之關聯推薦，以客觀系統指標及主觀滿意度來驗證。最後發現使用推薦機制前與後的「系統指標增加率」比較中，有學習風格關聯推薦是增加最多，減少最小的。使用推薦機制後，三組在系統指標「閱讀文章數」、「單次閱讀文章數」及「相關文章推薦點閱率」是有顯著差異的。主觀滿意度中「想要的」、「容易使用」、「整體滿意度」項目上有顯著差異，證明加入學習風格之關聯推薦是有效的，並且挖掘四種學習風格學習者的不同瀏覽模式。
關鍵字：學習風格、關聯規則、資料探勘、推薦機制

關鍵字

推薦機制 ； 學習風格 ； 資料探勘 ； 關聯規則

並列摘要

Abstract
Compared with traditional learning, E-learning model has some characteristics, such as convenience, the ignorance of distance, the reduced cost, and the enhancement of teaching quality. The researches about personalization on E-learning usually provide information technology, but ignore that the learning style is also an important factor and effects learning style. So this research devises “the learning style associate recommend mechanism” and provides acceptable leaning resources.
First, we use Kolb’s Learning Style Inventory to analyze leaning style. All learners are separated into four learning styles, and then they studied in the internet web. Next, we make an analysis of related rule for different browsing records of learners and conclude four browsing behavior of learning styles. Final, after comparing browsing records of learners with the rules, we recommend learners to skim the unread and relational articles.
The effect of “the learning style association recommend mechanism” can be proved by the following experiment. The learning resource including five subjects and fifty articles is the Reading Part of Intermediate level of GEPT (General English Proficiency Test). There are 117 testing subjects divided into 3 groups. They are freshmen and sophomores in Chung Yuan Christian University. Group A is non-recommend. Group B is association recommend. Group C is the learning style association recommend. It is proved by the objective system-indicators and subjective satisfaction.
From the result of experiment, rate of increasing system-indicators, compare use recommend mechanism before and after, in the learning style association recommend are increase most and reduce the minimum one. After we use recommend mechanism, there is a big difference between three groups in the items of system indicators “quantity of reading”, “quantity of single reading”, and “hitting ratio of relevant articles”. They are also different between the items of subjective satisfaction “want”, “easy using” and “integral satisfaction”. According to the results, we can dig not only the usefulness of “the learning style association recommend mechanism”, but also different browsing model of four learning styles.
Keywords: learning style, association rule, data mining, recommend mechanism

並列關鍵字

recommend mechanism ； learning style ； data mining ； association rule

參考文獻

21. 歐陽崇榮 (民89)，圖書館自動化系統評估模式之建立與應用，國立中央大學資訊管理研究所博士論文。

22. 蔣月美 (民91)，彈性能力分班在國小英語教學之實施--台北市龍安國小個案研究，國立台北師範學院兒童英語教育研究所，http://www.laes.tp.edu.tw/chiangym/essay.htm。

30. 韓?福 (民91)，應用協同過濾機制於商機撮合電子報之研究 –以食品業電子市集，中原大學資訊管理研究所碩士論文。

1. Baroudi, J.J., and Orlikowski, W.J. (1988), “A Short-Form Measure of User InformationSatisfaction: A Psychometric Evaluation and Note on Use,” Journal of Management Information Systems, 4(4), p. 44-59.

4. Chapelle, C. & Roberts, C. (1986), “Ambiguity tolerance and field independence as predictors in English as a second language,” Language Learning, 36(1), p.27-45.