Performance Comparison of Simple Linear Regression and Single Exponential Smoothing Methods for Forecasting Nickel Production at PT. Jaga Aman Sejahtera

Forecasting nickel production is a crucial step in the planning and operational decision-making process at PT. Jaga Aman Sejahtera. This study compares the performance of two forecasting methods, namely Simple Linear Regression and Single Exponential Smoothing, in predicting nickel production volumes based on the number of active machines. The evaluation results indicate that Simple Linear Regression has a higher level of accuracy, with lower MAE and MAPE values compared to the exponential smoothing method. Meanwhile, the performance of Single Exponential Smoothing is highly influenced by the selection of the alpha value, with prediction errors increasing at higher alpha values. Therefore, Simple Linear Regression is more recommended for forecasting nickel production, as it provides more stable and accurate predictions.