<![CDATA[Abstract: This research develops a web-based system to detect the similarity of thesis titles using the Modified Cosine Similarity method and TF-IDF weighting. This system helps students in evaluating the similarity of titles automatically, so as to avoid plagiarism and increase the originality of research. The process used includes text preprocessing (case folding, tokenizing, stopword removal, stemming), TF-IDF calculation for word weighting, and the use of Modified Cosine Similarity to measure the level of similarity between titles. The test results show that the system is able to identify the similarity of titles with a 100% recall rate, where titles with >70% similarity need to be revised, 31%-70% similarity can be clarified by adding words, and ≤30% similarity is potentially accepted as an original title. With this implementation, it is expected that students can more easily determine a thesis title that is unique and in accordance with academic standards.Abstrak: Penelitian ini mengembangkan sistem berbasis web untuk mendeteksi kemiripan judul skripsi menggunakan metode Modified Cosine Similarity dan pembobotan TF-IDF. Sistem ini membantu mahasiswa dalam mengevaluasi kemiripan judul secara otomatis, sehingga dapat menghindari plagiarisme dan meningkatkan orisinalitas penelitian. Proses yang digunakan meliputi text preprocessing (case folding, tokenizing, stopword removal, stemming), perhitungan TF-IDF untuk pembobotan kata, dan penggunaan Modified Cosine Similarity untuk mengukur tingkat kemiripan antar judul. Hasil pengujian menunjukkan bahwa sistem mampu mengidentifikasi kemiripan judul dengan tingkat recall 100%, di mana judul dengan kemiripan >70% perlu direvisi, kemiripan 31%-70% dapat diperjelas dengan penambahan kata, dan kemiripan ≤30% berpotensi diterima sebagai judul orisinil. Dengan implementasi ini, diharapkan mahasiswa dapat lebih mudah menentukan judul skripsi yang unik dan sesuai dengan standar akademik.]]>
Copyrights © 2025
