<![CDATA[The rapid growth of crypto assets and the variety of Centralized Exchange (CEX) platforms make it difficult for traders to choose a platform that fits their preferences. This research aims to model a recommendation system for CEX platforms using Collaborative Filtering. User rating data for several CEX (Binance, Bybit, Bitget, Tokocrypto, Indodax) were collected via questionnaire. The K-Nearest Neighbors With Means (KNN With Means) method with cosine similarity is used to predict ratings based on the similarity of preferences between users. The model was trained and tested with a 75:25 train-test split. Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) were used as evaluation metrics. Test results show low MAE and RMSE values (around below 1.0 on a 1–5 rating scale), indicating that the recommendations generated are quite accurate. It can be concluded that the Collaborative Filtering approach is effective in recommending CEX platforms according to user needs. This recommendation system is expected to assist traders – especially beginners – in choosing the right exchange more objectively.]]>
Copyrights © 2025
