IMPLEMENTASI SISTEM REKOMENDASI CONTENT-BASED FILTERING  UNTUK REPOSITORI SKRIPSI TEKNIK INFORMATIKA MENGGUNAKAN  COSINE SIMILARITY DAN JACCARD INDEX

Authors

Kata Kunci: Content-Based Filtering, Cosine Similarity, Jaccard Index, Repositori Skripsi, Sistem Rekomendasi, Universitas Malikussaleh.

The Informatics Engineering Department at Universitas Malikussaleh often faces challenges in assisting students

to find thesis topics that align with their interests and academic skills. This issue is exacerbated by the lack of an

effective thesis repository, leading to title duplication and inefficiency in topic selection. This study aims to

develop a recommendation system based on content-based filtering, utilizing Cosine Similarity and Jaccard Index

as methods to identify similarities between thesis documents. The system is designed to help students choose thesis

topics relevant to their interests while improving the department's efficiency in managing the thesis repository.

The method includes collecting thesis data from the department’s repository and web scraping, followed by data

preprocessing techniques such as tokenization, stop-word removal, and stemming. The system evaluation

measured precision, recall, and F1-score, with results showing a precision of 80% and a recall of 60%. This

research demonstrates that the system improves the thesis topic selection process, although further improvements

in accuracy and recommendation variety are still needed.