<![CDATA[The growth of e-commerce in the fashion industry has significantly increased the number of digital products offered to consumers. This situation leads to an information overload problem, where users find it difficult to choose products that match their preferences. To address this issue, a recommendation system is needed to filter information and provide relevant product suggestions. This study aims to develop a fashion product recommendation system based on user ratings using the Singular Value Decomposition (SVD) method. The data used is secondary data sourced from the Kaggle platform, containing user interactions with various fashion products. The research process includes data collection, preprocessing, rating matrix construction, data decomposition, and model performance evaluation. The evaluation results show a Mean Absolute Error of 0.47 and a Root Mean Squared Error of 0.67, indicating a relatively low prediction error. The system also successfully recommends items with high predicted ratings, such as a rating of 4.30, demonstrating strong relevance to user preferences. These findings confirm that the applied method is effective in building an accurate and relevant recommendation system that assists users in making fashion product purchase decisions more efficiently.]]>
Copyrights © 2025
