Analysis of Rainfall Forecasting Using the Holt-Winters Approach: Additive and Multiplicative in Pekanbaru

Rainfall plays an important role in various sectors, such as agriculture, water resource management, and disaster mitigation. However, rainfall patterns in many regions of Indonesia tend to fluctuate and change from year to year, making accurate forecasting increasingly necessary. Several previous studies conducted in Sumbawa Regency, North Padang Lawas Regency, Ambon City, and Pasuruan have shown that the Holt–Winters Exponential Smoothing method is effective in forecasting monthly rainfall because it is able to capture trend and seasonal patterns in the data, resulting in forecasts that are closer to actual conditions. This study aims to analyze and forecast monthly rainfall in Pekanbaru City using the Holt–Winters Exponential Smoothing method for the 2019–2024 period. The data used consist of monthly rainfall data analyzed using additive and multiplicative Holt–Winters models. The analysis results show that the rainfall data exhibit a clear and relatively stable seasonal pattern from year to year, indicating that the Holt–Winters method is appropriate for use. Model performance evaluation using Mean Absolute Error (MAE) shows that the additive Holt–Winters model produces a smaller MAE value than the multiplicative model. This indicates that the additive model provides better forecasting accuracy in modeling monthly rainfall in Pekanbaru City.