Unleashing the Power of Time-Series Method: Iluminating Indonesia's Open Unemployment Rate Based on Educational Attainment

This study endeavors to project the Unemployment Rate in Indonesia based on Highest Education categories for the years 2021 and 2022. The impetus for this research stems from the persistent high unemployment rate in the country, signaling a need for heightened efforts in job creation from both governmental and private sectors. Employing time-series data obtained from the Badan Pusat Statistik website (2006–2020), the research utilizes established methodologies, such as Exponential Smoothing and ARIMA, within the CRISP-DM research flow framework and RStudio software. Model selection hinges on the Mean Absolute Percentage Error (MAPE), with Exponential Smoothing emerging as the optimal model. The analysis reveals Exponential Smoothing as the most apt time-series model for forecasting unemployment based on highest education categories in 2021 and 2022. The model uncovers nuanced variations in unemployment patterns, with a discernible seasonal trend characterizing average unemployment rates over the next two years. Specifically, unemployment in SMP (junior high school) remains relatively constant, while S1 (University) exhibits an upward trend. These findings underscore the intricate dynamics influencing unemployment forecasts in Indonesia, emphasizing the consequential role of educational disparities in shaping labor market outcomes.