基於本體論之程式設計除錯測驗與迷思概念之探勘

An Ontology-Based Programming Assessment by Misconception Mining

范揚棋(fan yang ch'i)

指導教授 ： 曾憲雄

亞洲大學/資訊學院/資訊科學與應用學系碩士班/碩士(2007年)

摘要

解決學生在程式設計時碰到的問題是一件具挑戰且重要的議題，根據我們的 觀察，學生在寫程式時常常會有一些迷思的概念(包含忘記與不熟)，導致寫出不 正確的程式，像是少加符號的錯誤或是誤用符號的錯誤。為了解決此問題，本研 究提出使用本體論來輔助描述學生錯誤的概念，並進一步使用資料探勘的方法以 找出錯誤概念間的關係。另外我們提出一套基於本體論的程式設計迷思概念分析 架構，以達到我們想要找出學生程式設計概念與錯誤型態的問題。此架構包含三 個階段：第一階段，建構以框架為主的程式概念階層階段；第二階段，產生以除 錯為主的試題階段；第三階段，迷思概念分析階段。在我們建構程式概念本體論 後，我們於階段一可以將所謂的程式概念以框架的方式來表達。同時，研究中並 建構了程式錯誤型態本體論，利用框架來詳細描述各種錯誤的型態。經過關聯式 分析探勘，可依概念間的關係產生學習概念圖，以提供老師在教學策略上的建 議。由實際施測於五十位學生後，我們發現本研究成果對於改善學習成效及教學 策略都有相當的助益。

關鍵字

學習程式設計、迷思概念、程式概念、先備知識、本體論、框架、資料 探勘

並列摘要

It is an important and challenging issue to overcome the problem within the
student learning in programming. From the observations, there are many programming
misconceptions occur when the student writing their programs. Several bugs such as
missing token or misuse token frequently occur. We want to find out whether the
student incautious or they have misconception about the programming concepts, or
whether there are some prerequisite concepts they didn’t learn properly. We, then,
propose the idea to find out what reason does behind this fact. We propose an
Ontology-Based Programming Misconception Analyzing Scheme (OBPMAS) which
has the following three phases to reach our goal: (1) Phase 1: Frame-Based Program
Concept Hierarchy Construction, (2) Phase 2: Debugging Test Item Generation and (3)
Phase 3: Misconception Analyzer. After Program Statement Ontology has been
constructed, we model the concept with each frame above in Frame-Based Program
Concept Hierarchy Construction phase. We construct the Buggy Pattern Ontology
(BPO) containing about many error types, where is also use frame to describe in detail
about BPO. We analyze the answer sheet by using association rule mining to construct
the concept map. From that answer sheet, we can predict what misconceptions the
student may have. Then we can obtain the feedback as the strategy for the teacher to
improve student learning in programming course. Experiments we did involved about
fifty persons (i.e., the student). The result show us the research we did that help
improve learning effects and learning strategy.

並列關鍵字

Programming Learning, Misconception, Programming Concept, Prerequisite Concept, Ontology, Frame, Data Mining

參考文獻

[2] Crews, T. & Ziegler, U. (n.d.), The Flowchart Interpreter for Introductory Programming

[4] Garner, S., “Learning Resources and Tools to Aid Novice learners Learn

Tutoring System for Programmable Logic Controller,” 33rd ASEE/IEEE Frontiers in

Education Conference, Boulder, CO, Nov., 2003.

[6] Rowe, G. & Thorburn, G., “VINCE - an On-Line Tool for Teaching Introductory