Perancangan Fitur Deteksi Kemiripan Dokumen Jawaban Tugas Mahasiswa pada Sistem Manajemen Pembelajaran dengan Metode K-Shingling dan Cosine Similarity

Abstract

Technology development can be used to support the teaching and learning process so it can runs more effectively and efficiently, and one of the applications that can be used to support this process is the Learning Management System (LMS). LMS provides an integrated platform starting from delivering learning materials to students, and evaluating learning process. The evaluation method that can be done at the LMS is giving assignments to students, and students can do submission by uploading their documents assignment answer. The common problem is many students do plagiarism by duplicate the answers of other students. To overcome these problems, this study will implement feature in Course Assignment Menu in LMS to detect the similarity of student assignment document using K-Shingling and Cosine Similarity methods. The K-Shingling algorithm is used to form shingles or word fragments based from all text in the task answer document that has gone through the pre processing stage, then each document have shingle value vector which will be used to compare the similarity values between documents using Cosine Similarity. The results obtained in this study is the developed feature is capable to detecting the similarity percentage of the assignment answer documents from each student, and from that percentage value it can indicate whether the students assignment answers is plagiarism to other students assignment answers.

Keyword : LMS; plagiarism; K-Shingling; Cosine Similarity