<![CDATA[Introduction: This study aims to analyze the use of machine learning in predicting the number of customers for Amanah products at PT. Pegadaian CPS Pangkajene. The issues examined include the application of machine learning, the selection of the most effective algorithm, and factors affecting prediction accuracy. Methods: The study uses a quantitative approach, drawing on historical customer data over five years. The three machine learning models compared are SARIMA, Random Forest, and Linear Regression. Results: The SARIMA model achieved the best performance, with a Mean Absolute Error (MAE) of 2.61%, a Root Mean Square Error (RMSE) of 2.94%, and a Mean Absolute Percentage Error (MAPE) of 8.58%. This model is the most accurate in predicting trends and seasonal patterns in customer numbers. The Linear Regression model showed less accurate results, while Random Forest excelled in reliability but was still less accurate than SARIMA.Conclusion and suggestion: The application of machine learning, especially SARIMA, can be a strategic tool for business planning and decision-making in the Islamic finance sector, providing data-based recommendations to improve the performance of Amanah products. Keywords: Amanah Product, Customer Prediction, Machine learning, Marketing Strategy, PT. Pegadaian, SARIMA]]>
Copyrights © 2026
