最鄰近法應用之自我解釋分類機制

The Mechanisms of Applying K-Nearest Neighbor Algorithm for Classifying Self-?Explanations

簡瑞泉(Ruei-Chuan Chien)

指導教授 ： 周志岳

元智大學/資訊學院/生物與醫學資訊碩士學位學程/碩士(2012年)

https://doi.org/10.6838/YZU.2012.00195

摘要

自我解釋是一種自我建構式學習活動，該活動要求學生在研讀教材產生自我解釋來釐清和詮釋教材的內容，以及自我監測對教材的了解狀況。學生會產生許多類型的自我解釋，例如重述教材內容、意譯、連結性推論、領域先備知識推論、邏輯性推論、正面自我監測、負面自我監測等類型。通常要了解學生所產生自我解釋的類型，需要以人工的方式來分類自我解釋。本研究利用資訊探勘相關的技術提供五種不同的自我解釋分類機制用來自動分類學生的自我解釋，包括意譯、連結性推論、邏輯性推論、正面自我監測、負面自我監測五種類型。本研究先透過分詞技術分解學生的自我解釋，再使用TF-IDF權重計算方法找出學生自我解釋中的關鍵字並且給予特徵值，接著配合餘弦相似度建立向量空間模型和判斷自我解釋之間的相似性，最後藉由傳統K最鄰近法、適應性K最鄰近法、納入教材內容之適應性K最鄰近法、納入教材內容並去除通用字詞之適應性K最鄰近法、以及採多重分類之適應性K最鄰近等五種不同分類機制之來分類出自我解釋的種類。評估結果是以採多重分類之適應性K最鄰近法有最佳的分類準確性，在正面自我監測與負面自我監測的準確性可達到九成，而連接性推論、邏輯性推論與意譯的準確性也比其它類的分類機制較佳，整體的正確率則可達到七成。

關鍵字

自我解釋 ； TF-IDF ； K最鄰近法

並列摘要

Self-explaining is a self-constructive learning activity, which engages students in clarifying and explaining the content and self-monitoring their understanding of the content. Students may generate many kinds of self-explanations, such as re-reading, paraphrase, bridging inference, prior knowledge inference, logic inference, self-monitoring. In general, student explanations need to be classified by human experts, and the classification is time-consuming and labor-intensive. This study applies data mining techniques to automatically classify student explanations. This study adopts vector space model to represent student explanations and applies K-nearest neighbor mechanism to classify student explanations. This study investigates and compares five K-nearest neighbor classification mechanisms: traditional K-Nearest Neighbor,　adaptive k-Nearest Neighbor,　adaptive k-Nearest Neighbor which includes content for self-explaining,　adaptive k-Nearest Neighbor which includes content for self-explaining and excludes common words, adaptive k-Nearest Neighbor with multiple classification. The evaluation results show that adaptive k-Nearest Neighbor with multiple classification has best classification correctness. The correctness of classifying self-monitoring is about 90% and total correctness of classification is about 70%.

並列關鍵字

Self-explaining ； TF-IDF ； K-Nearest Neighbor

參考文獻

參考文獻

Aleven, V., Pinkwart, N., Ashley, K., & Lynch, C.(2006), “Supporting Self-explanation of Argument Transcripts: Specific v. Generic Prompts”, Proceedings of the Workshop on Intelligent Tutoring System for Ill-Defined Domains at the 8th International Conference on Intelligent Tutoring System,47-55.

Chi, M. T. H., Bassok, M., Lewis, M., Reimann, P., & Glaser, R. (1989). Self-explanations: How students study and use examples in learning to solve problems. Cognitive Science, 13, 145-182.

Chi, M. T. H., de Leeuw, N., Chiu, M.H., & LaVancher, C. (1994). Eliciting self-explanations improves understanding. Cognitive Science, 18, 439-477

Chi, M. T. H. (2000). Self-explaining expository texts: The dual processes of generating inferences and repairing mental models. In R. Glaser (Ed.), Advances in Instructional Psychology, Hillsdale, NJ: Lawrence Erlbaum Associates. 161-238.