Sistem Rekomendasi Berbasis Collaborative Filtering Pada E-Commerce

Abstract

The rapid development of e-commerce causes information overload, where users have difficulty finding suitable products amidst the many choices. Recommendation systems are becoming a key component for improving user experience and driving sales. This research aims to design and implement a product recommendation system in e-commerce using the Collaborative Filtering method (both User-based and Item-based). This method works by analyzing user behavior patterns, such as transaction history, ratings, or clicks, to look for similarities between users or between items. The Cosine Similarity technique is used to measure similarity, while k-Nearest Neighbor (KNN) is applied to find the nearest neighbors to produce predictions. This system is designed to overcome the problem of data sparsity and provide personalized recommendations. System evaluation is carried out using the Mean Absolute Error (MAE) or Root Mean Squared Error (RMSE) metric to measure the level of prediction accuracy. The research results show that a recommendation system based on Collaborative Filtering is able to produce relevant product recommendations and increase the effectiveness of marketing strategies on e-commerce platforms.