Implementasi Metode Triple Exponential Smoothing pada Sistem Peramalan Produksi Pertanian Kabupaten Bojonegoro

Abstract

Agriculture is one of the essential sectors in Indonesia that contributes to economic growth, fulfills food and industrial needs, and absorbs a large labor force. To support decision-making in the agricultural sector, production forecasting plays an important role. This study aims to implement the Triple Exponential Smoothing method in the agricultural production forecasting system of Bojonegoro Regency and to evaluate the accuracy of the method. The forecasting method used in this study is the Triple Exponential Smoothing with a multiplicative model, which was integrated into a PHP-based system. The results show that the developed system is capable of processing data and producing forecasts that align with manual calculations using Microsoft Excel. However, the accuracy test using the Mean Absolute Percentage Error (MAPE) resulted in a value of 126%, which is categorized as poor. The high MAPE value is caused by unstable data fluctuations and irregular seasonal patterns, especially in March, April, and June, which consistently show significant production spikes each year. Therefore, the Triple Exponential Smoothing method is considered less effective in modeling agricultural production data with unstable seasonal patterns