Article Reviewer Recommendation System Using Euclidean Distance Similarity with Content-Based Collaborative Filtering (Case Study: ICVEE)

Abstract

The growth of research publications in academic environments has resulted in large volumes of unstructured data, particularly in the form of article titles and abstracts. However, the majority of educational institutions still manage these resources manually, without optimizing them for academic decision-making. This study proposes an article reviewer recommendation system using a content-based filtering method with TF-IDF for text representation and Euclidean Distance as the similarity measure. Reviewer profiles are constructed based on previously reviewed articles. A new article is represented as a vector and compared against reviewer profiles to determine relevance. The system was evaluated using 20 articles as ground truth. Results show that the Euclidean Distance approach outperformed Cosine Similarity, achieving an accuracy of 55%, precision of 0.2333, recall of 0.2121, and F1-score of 0.222. This study demonstrates the potential of content-based filtering in enhancing reviewer assignment efficiency for academic conferences such as ICVEE.