電腦適性化測驗雛型系統之研究

Abstract

電腦科技與網路技術的發展造成數位學習日益受到重視，國際組織也訂出相關標準(Sharable Content Object Reference Model, SCORM)給予一個交換資料的標準，有別於傳統教學方法，數位學習可以達到節省人力、時間等成本，適性化電腦輔助測驗(computerized-adaptive test, CAT)的提出也就是利用電腦的快速計算能力，針對學生的不同程度給予不同能力的測驗，目的在於準確的找出學生能力值。 本研究目的在電腦適性化測驗方面做探討，以概念關聯架構為主軸，發展出一套以國小數學為範圍的適性測驗雛型系統，為防止學生猜題的可能，系統中加入選題策略，能將猜題機率控制在一定值之下。系統除了測驗功能外，也提供試題編輯、系統管理及成效分析等線上功能，以有效的題目測出學生的失敗概念，進而可以給予補救教學建議，加上家長及教師的參與，可以立即對學生達到引導學習的效果。

With the development of computer science and network technology, e-learning plays a more important role than ever. SCORM which is an e-learning standard developed by ADL defines how the interoperability can be achieved. The computerized adaptive test (CAT) is to utilize the computing power of computer to precisely give the specified exam for the specified category of examinees, and find out what the student ability is. This paper is aimed to develop a prototype system of CAT based on Concept Relation Structure (CRS). To avoid from guessing we use item selection strategies to control the guessing probability smaller than the threshold. The proposed system also includes item edition, performance evaluation, and system management. After taking exam student can acquire the learning comments from the system, and the tutors and parents can guide the learner with the system advices.