본문 바로가기
  • Home

Assessment of exEyes' Recall using Bellon Reference Corpus as a Benchmark

  • Journal of Software Assessment and Valuation
  • Abbr : JSAV
  • 2015, 11(1), pp.31-39
  • Publisher : Korea Software Assessment and Valuation Society
  • Research Area : Engineering > Computer Science
  • Received : May 22, 2015
  • Accepted : June 20, 2015
  • Published : June 30, 2015

Sungha Choi 1 Doh, Kyung-Goo 1

1한양대학교

ABSTRACT

Copyrights for software source codes are given to developers. Korea Copyright Commission utilizes a clone-detection tool, exEyes, to find code clones that can be used to assess software plagiarism. This paper evaluates the recall of exEyes using Bellon Reference Corpus as a benchmark. Four open sources(cook and weltab in C, eclipse-ant and netbean-javadoc in Java) in Bellon Reference Corpus are selected as the benchmark. Among 10,055 clones in the corpus, exEyes' recall rate is 100% in clone type 1, 63% in clone type 2, and 34% in cone type 3. False negatives turn out to be mainly caused by ignoring the meaning of tokens when the comparison is made, and by setting the comparison be made line-by-line.

Citation status

* References for papers published after 2023 are currently being built.