基於資料探勘之程式設計迷思概念診斷
No Thumbnail Available
Date
2019
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
過去程式設計迷思概念相關研究常有研究樣本過少以及解讀較為主觀的問題,導致無法產出較一般化的研究結果,且無法將其作廣泛應用。這是因為過去的研究多以研究者對受試者的診斷為依據,除了較為主觀,亦須耗時去解讀與分析資料。本研究利用資料探勘於程式設計迷思概念探究,透過蒐集大量程式設計初學者的程式碼資料,探勘初學者若擁有迷思概念可能寫出的程式碼特徵,並設計診斷機制,根據學習者所撰寫的程式碼診斷其可能的迷思概念。本研究並將迷思概念探勘與診斷機制實作於學習平台上,以提供學習者即時的協助。研究參與者共1216人,包含527位大學生與689位高中生。參與者必須完成程式設計測驗,並接受抽樣訪談。本研究根據程式設計測驗的作答結果進行資料探勘,找出各程式設計迷思概念之程式碼症狀,並據此進行學習者程式設計迷思概念之診斷。透過資料探勘,我們亦發現不同的迷思概念間存在著關聯性:若學習者對for迴圈條件判斷的運作流程有迷思概念,則對if-else條件判斷的運作流程亦具有迷思概念。此外,在程式碼分群的結果分析中,我們額外發現幾種文獻上未提到的迷思概念。本研究之結果將提供教學者有效的教學改進依據,並能用以即時診斷學習者的程式設計迷思概念,以進行後續的矯正,進一步提升程式設計能力。
Most of the research related to the concept of the misconception programming in the past faced the same limitations, including too few research participants, only locking specific programming languages in research, and not being able to apply the results to a wide range of people. Most of these restrictions are caused by research methods. Most of the past research is based on the self-judgment of researchers as the basis for the research results of misconception. Therefore, it takes a considerable amount of time to organize and analyze the data. Therefore, the scope of the research data is limited. Most of the restrictions on the masses occur, and most of the programming tests or tools used for diagnosis only develop into specific programming languages, which limits the scope of application of the research results. Therefore, this study is different from the past research methods and uses data mining in the program. Programming misconception of data mining is to use algorithms to find similar clusters between feature vectors, and to find out the misconception types and misconception symptoms. Through interviews, the correctness of the misconception data mining results is confirmed, and the programming learning platform for the misconception diagnosis is developed. The results of the study are based on the three main research objectives, including the narrative of the misconception of programming and the corresponding symptoms of the program, as well as the diagnostic mechanism of the development of the minconception data mining, and pointing out the relationship between misconceptions. The misconception of programming is a topic of education that has been discussed for a long time. The misconception does cause learning problems in the programming to learners. This study is based on the established misconception, adding the factors of data mining, and obtaining research. In the design of misconception and the research unit is locked into process control. The results can be applied to a wide range of programming languages. Teachers and learners have substantial support in teaching and learning. They can obtain more appropriate teaching guidelines and learning corrections to enhance the effectiveness of programming learning.
Most of the research related to the concept of the misconception programming in the past faced the same limitations, including too few research participants, only locking specific programming languages in research, and not being able to apply the results to a wide range of people. Most of these restrictions are caused by research methods. Most of the past research is based on the self-judgment of researchers as the basis for the research results of misconception. Therefore, it takes a considerable amount of time to organize and analyze the data. Therefore, the scope of the research data is limited. Most of the restrictions on the masses occur, and most of the programming tests or tools used for diagnosis only develop into specific programming languages, which limits the scope of application of the research results. Therefore, this study is different from the past research methods and uses data mining in the program. Programming misconception of data mining is to use algorithms to find similar clusters between feature vectors, and to find out the misconception types and misconception symptoms. Through interviews, the correctness of the misconception data mining results is confirmed, and the programming learning platform for the misconception diagnosis is developed. The results of the study are based on the three main research objectives, including the narrative of the misconception of programming and the corresponding symptoms of the program, as well as the diagnostic mechanism of the development of the minconception data mining, and pointing out the relationship between misconceptions. The misconception of programming is a topic of education that has been discussed for a long time. The misconception does cause learning problems in the programming to learners. This study is based on the established misconception, adding the factors of data mining, and obtaining research. In the design of misconception and the research unit is locked into process control. The results can be applied to a wide range of programming languages. Teachers and learners have substantial support in teaching and learning. They can obtain more appropriate teaching guidelines and learning corrections to enhance the effectiveness of programming learning.
Description
Keywords
迷思概念診斷, 程式設計教學, 資料探勘, Misconception diagnosis, Programming teaching, Data exploration