結合順序理論與貝氏網路之學生概念診斷模型

Modeling students’ learning bugs and skills based on Bayesian network combined with order theory

透過以專家為觀點的方式建構貝氏網路是相當費時且費力的，這樣的窘境在當網路節點過於龐大時更顯而易見。因此，本研究的目的在於提出一個基於順序理論的學生知識結構建立演算法，並用以作為建立貝氏網路結構的一個基礎。此外並透過該演算法與貝氏網路診斷模型作結合，已期望可以快速建立以貝氏網路為基礎的學生概念診斷模型，來克服以往必須花費大量人力與物力的開發窘境。為了驗證該方法的有效性，本研究將以專家建立的貝氏網路結構作為校標，並透過實證研究的方式來進行評估。評估結果發現，所提出的新演算法所建立的學生知識結構並將之應用在概念診斷模型中，無論在演算速度或是學生錯誤概念與技能的辨識效果上，都遠比專家所建立的方式來得好。

Because eliciting Bayesian networks modeling students' learning bugs and skills from experts can be laborious and difficult procedure in the case of large networks, the goal of this paper is trying to develop an algorithm based on order theory for structure learning of Bayesian network from data. And using proposed methodology constructs the structure of Bayesian networks modeling students' learning bugs and skills and applying to diagnosis students' bugs and skills..The evaluation indicates that the procedure for constructing structure is faster and the classification accuracy is better in comparison with the structure constructed by experts' prospective.