- 學位論文
貝氏網路適性診斷測驗暨學習系統融入國中英語時態教學
Bayesian Network Adaptive Diagnostic Test and Learning System Integrated into Teaching of English Tenses in Junior High School
摘要
中文摘要 本研究採用貝氏網路適性診斷測驗系統編製一套國中英語時態診斷測驗試題，分析學生在英語時態方面常犯的錯誤類型，以此編製電腦輔助教學與補救教學媒體，並應用於教學實驗中。研究結果發現： 1. 貝氏網路的錯誤類型及子技能的辨識率平均值為92.32%，進行電腦化適性測驗後，試題的省題率平均值為16.5%，且縮短了施測時間。在試題預測精準度上，其預測精準度平均值為94.4%，顯示其預測精準度相當不錯。 2. 在教學成效比較分析方面，在電腦輔助教學模式與傳統教學模式下，實驗組學生與對照組學生的成績未達顯著差異。 3. 在電腦化補救教學方面，不論高分組、中分組或低分組學生都有進步，尤其以低分組學生進步幅度最大，顯示電腦化補救教學有助提高學習效果。 4. 根據學生學習回饋意見統計表發現，學生認為電腦輔助教學及補救教學的教材及模式對他們的學習有幫助，而在英語時態的難度認知方面，學生認為現在完成式是困難的，只有少數人認為現在進行式是困難的，研究者從這些意見中獲得相當大的回饋，可作為日後教學內容及教學方式的修正依據。 關鍵詞：貝氏網路、時態、子技能、錯誤類型、電腦化適性診斷測驗
並列摘要
Abstract This study attemps to establish the assessment of English tenses to computer adaptive diagnostic test based on Bayesian Network and analyze the error types that students often make in English tenses. According to the error types, apply computer-assisted teaching media and computer remedial teaching media of the establishment to the teaching experiment. The results are as follows. 1. Bayesian network’s error patterns and sub-skills of the average recognition rate was 92.32 %, after carrying on the computerized adaptive test, the average saving rate of test questions reached 16.5%, and reduced testing time. Prediction accuracy on the test questions, the average prediction accuracy was 94.4%, indicating that the prediction accuracy was quite good. 2. In comparative analysis of teaching effectiveness, in the computerized adaptive testing way and traditional teaching way, the achievement of the students of experiment group and control group students has not reached the significant difference. 3. In computer-based remedial instruction, no matter the high-level group, the middle-level group or the lower-level group, the students had good progress, especially in the lower-level group. It reveals computer-based remedial instruction help rase learning effect. 4. Feedback tables based learning found that students thought that computer-assisted teaching and remedial teaching materials and models was helpful to their learning, but when the cognitive respect of difficulty in English tenses, the students thought that the present perfect tense was difficult, only a few people thought that the present progressive tense was difficult, the researchers obtained from these observations considerable feedback, can be used as teaching content and teaching methods in future amendments to the basis. Key words: Bayesian Networks, tenses, sub-skills, error patterns, Computerized Adaptive Testing