Forecasting the Number of Tourist Visits to Semarang City Using the Holt-Winters Multiplicative and Fourier Series Analysis (FSA) Methods

Abstract

Forecasting is a crucial process in strategic planning, including in the tourism sector, which is highly influenced by seasonal fluctuations and external factors. In Semarang City, the number of tourist visits has experienced significant changes in recent years, particularly due to the COVID-19 pandemic, which caused a sharp decline in 2020 and 2021. Nevertheless, a post-pandemic upward trend in visits has begun to emerge. To support adaptive, data-driven planning, accurate forecasting methods capable of capturing both trend and seasonal patterns are essential. This study aims to compare the Holt-Winters Multiplicative method and Fourier Series Analysis (FSA) in forecasting tourist visits to Semarang City. The data used consist of monthly tourist visit records from 2017 to 2024, obtained from the Semarang City Office of Culture and Tourism. Forecast accuracy was evaluated using Mean Absolute Error (MAE) and Symmetric Mean Absolute Percentage Error (SMAPE). The results show that the Holt-Winters Multiplicative method produced a lower MAE of 44,498.31 and SMAPE of 6.76%, compared to the Fourier Series Analysis method. Therefore, the Holt-Winters Multiplicative method is considered more accurate in forecasting tourist visits. This study is expected to contribute to more effective tourism management and development strategies at the regional level.