<![CDATA[The adoption of information technology in predictive systems is increasingly popular in the agricultural export sector, leveraging historical data analysis to forecast global market fluctuations and optimize supply chains for commodities such as bananas, coffee, and coconuts. Exporters in North Sumatra face challenges related to fluctuating export volumes to Malaysia, influenced by seasonal factors, international price changes, weather dependency, and a lack of accurate data. This results in supply imbalances, economic losses, and difficulties in strategic planning. This research offers a solution by employing the Autoregressive Integrated Moving Average (ARIMA) method in the development of a web-based system to address these issues. ARIMA is a statistical time series model that combines autoregressive (AR) components for dependencies on previous values, integrated (I) components to handle non-stationarity through differencing, and moving average (MA) components to predict the influence of past errors; its seasonal variant (SARIMA) is applied to capture monthly harvest cycle patterns. The developed solution involves processing historical export data from 2019–2024 sourced from the Central Bureau of Statistics (BPS) via a Python Flask API with an automated ARIMA approach, integrated into a PHP CodeIgniter 4 web platform, providing interactive visualizations, real-time data updates, and easy user access. The expected outcomes from this system are more accurate export volume predictions, with a MAPE of approximately 19.02% and MAE of 180,755.28 tons on 2024 test data for bananas as a representative sample, which can support strategic decision-making, production efficiency, and enhanced competitiveness for banana, coffee, and coconut exports from North Sumatra to Malaysia.]]>
Copyrights © 2026
