Code analyzer for an online course management system

  • Authors:
  • Jong Yih Kuo;Fu Chu Huang

  • Affiliations:
  • Department of Computer Science and Information Engineering, National Taipei University of Technology, Taipei 106, Taiwan;Department of Computer Science and Information Engineering, National Taipei University of Technology, Taipei 106, Taiwan

  • Venue:
  • Journal of Systems and Software
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The online course management system (OCMS) assists online instruction in various aspects, including testing, course discussion, assignment submission, and assignment grading. This paper proposes a plagiarism detection system whose design is integrated with an OCMS. Online assignment submission is prone to easy plagiarism, which can seriously influence the quality of learning. In the past, plagiarism was detected manually, making it very time-consuming. This research thus focuses on developing a system involving code standardization, textual analysis, structural analysis, and variable analysis for evaluating and comparing programming codes. An agent system serves as a daemon to analyze the program codes for OCMS. For textual analysis, the Fingerprinting Algorithm was used for text comparison. Structurally, a formal algebraic expression and a dynamic control structure tree (DCS Tree) were utilized to rebuild and evaluate the program structure. For variables, not only the relevant information for each variable was recorded, but also the programming structure was analyzed where the variables are positioned. By applying a similarity measuring method, a similarity value was produced for each program in the three aspects mentioned above. This research implements an Online Detection Plagiarism System (ODPS) providing a web-based user interface. This system can be applied independently for assignment analysis of Java programs. After three comparison experiments with other researches, the results demonstrated the ODPS has many advantages and good performance. Meanwhile, a combined approach is proven that it is better than a single approach for source codes of various styles.