Penentuan Metode Forecasting untuk Prediksi Stok Telur Bebek di Yamuna Towuti

Abstract

A common problem faced by the poultry sector in Indonesia is the high demand and supply of eggs, which has an impact on price instability and the risk of losses for producers and distributors. This condition is also experienced by Yamuna Towuti, a duck egg distribution business in East Luwu, which has difficulty in maintaining consistent stock availability. A specific problem that has arisen is the lack of an accurate forecasting method to predict future demand for duck eggs, which often leads to mismatches. This study aims to determine the most accurate forecasting method for predicting duck egg demand by comparing three approaches, namely Weighted Moving Average (WMA), Exponential Smoothing, and Linear Regression/Least Squares. The data used is secondary data in the form of historical records of duck egg demand in 2024, which was then processed using POM software. The results show that the Linear Regression/Least Squares method has the best accuracy with the smallest MSE value of 489.792 and MAPE of 7.645%. Thus, linear regression can be recommended as the appropriate forecasting method for Yamuna Towuti to support decision-making in managing duck egg stocks and distribution more optimally.