Deteksi Penjiplakan Kode Program C dengan Metode Bisecting KMeans pada Kode yang Memiliki Makro
Abstract
Submission of programming tasks in digital form is vulnerable to plagiarism because the number of submitted codes can reach hundreds every week, making the process of plagiarism checking manually infeasible. Therefore we need a system that can detect plagiarism program code automatically. This has been done in previous studies using the K-means and bisecting K-means, but previous research assumes no preprocessor directive in the program code. Meanwhile, the use of preprocessor directive can change the structure of the program code that could affect the ability of the system to detect plagiarism. The purpose of this study is to detect codes that uses macro and compare the results obtained with previous research. The results obtained showed an increased accuracy in detecting the whole program code from 85 % to 89 %. Testing for each type of program code result in an increase of accuracy from 20%-40% to 80%.
Collections
- UT - Computer Science [2235]