<![CDATA[The beauty industry in Indonesia is experiencing rapid growth, particularly through e-commerce platforms like Tokopedia. Many businesses still rely on intuition for product management, including decisions related to stock and pricing. This study develops a machine learning-based classification model to identify beauty products with high sales potential on Tokopedia, considering factors such as price, rating, review count, and stock availability. Ten classification algorithms are applied, including Naive Bayes, SVM, K-Nearest Neighbors, Decision Tree, Random Forest, XGBoost, LightGBM, CatBoost, Extra Trees, and Multi-Layer Perceptron (MLP). The data is processed using Python on Google Colab. The results show that ensemble algorithms, particularly Random Forest, LightGBM, and Extra Trees, provide prediction accuracy above 91% and are highly effective in predicting best-selling products. Based on this model, businesses can optimize stock and pricing management to ensure that best-selling products are always available, thereby improving operational efficiency in a highly competitive market. This research offers a data-driven solution for more strategic and evidence-based product management on e-commerce platforms.]]>
Copyrights © 2025
