本研究以國中數學領域七年級之「一元一次方程式」單元為例,以貝氏網路理論基礎建立一套適用於本單元的電腦適性教材及電腦適性診斷測驗試題,再進行教學及補救教學成效評估。 研究結果發現: 1. 將紙筆診斷測驗轉換成為電腦適性化診斷測驗之後,適性化診斷測驗的施測的辨識率約91%,而適性測驗平均施測約26題,與傳統紙筆測驗相比,平均可以節省約8題,縮短了施測的時間。而在試題預測精準度上,試題、錯誤類型及技能的預測準確率在90%以上。 2. 在電腦化教學模式及傳統教學模式比較,不管經過教學、補救教學或學後保留,實驗組學生和對照組學生的成績達到顯著性差異,且實驗組的平均分數大於對照組學生的平均分數 3. 在電腦化補救教學模式下,子技能的達成率有增加的趨勢,而錯誤類型的發生率有降低的趨勢,其中,在「一元一次式子單元」低分組的學生及「解一元一次方程式子單元」中分組及低分組學生在使用本套教材後,進步有顯著的差異。 4. 在學習回饋問卷分析中,首先,有八成以上的學生在使用利用電腦化教材,提供較佳的動機;其次,學生在學習的難度認知是隨著基本概念認知、式子運算、應用問題的概念層次而提高的;最後,約有八成的學生認為電腦化教學是有幫助的。 因此,本研究所建置適性診斷系統及以貝氏網路為基礎的電腦化教材,的確可以達到良好的效果。
The main purpose of the study is to establish the materials of computer adaptive and the items of computer adaptive diagnostic test based on Bayesian Networks by the one-variable linear equation unit of seventh grades. Use the materials to teach and evaluate the effects of remedial instruction. The results are as follows. 1. After taking computer adaptive diagnostic test, the compare rate can reach around 91%. The computer adaptive diagnostic test economized 9 items on average. For items predictable accuracy, items, bug type and skills can reach above 90%. 2. By comparison with the traditional teaching model, the computer teaching model shows that experimental groups’ grades and control groups’ grades have significant differences. In addition the average score of experimental groups are higher than control groups. 3. Of remedial teaching in the computer model, the rates of sub-skills reach an increasing trend, while the incidences of error types reduce the trend. It shows that there is significant progress on the low group grades students of one-variable polynomial subunit and the low and middle group grades students of one-variable linear equation subunit using the computerized teaching materials. 4. First, in learning feedback questionnaire analysis, over eighty percent students using the computerized teaching materials have better motivation for learning. Second, the student difficulty cognitive of learning increases with the basic concept, polynomial operation, and application problem. At last, over eighty percent students think that it is helpful to use computerized teaching. Therefore, the computerized adaptive dagnostic testing system and computerized teaching materials based on Bayesian Networks proposed in this study can achieve good results.