FORECASTING EXPORTS IN ALBANIA

Time series forecasting is a significant domain, in fields such as economics, finance, and weather forecasting. The paper assesses time series forecasting that project forward-looking values based on on historical data. The common models used to capture the trend, seasonality, and anomalies within the time series data are ARIMA, SARIMA, and exponential smoothing. Forecast accuracy depends on the model choice and fine-tuning. Validation and continuous refinement are necessary for relia-bility and adaptability to changing conditions. Time series forecasting is a valuable tool for decision-making, resource allocation, and planning. An ARIMA model forecasted Albanian exports, providing reliable forecasts for the near future and prioritizing recent observations over past data. One draw-back is that new data were not applied to different models for improving future forecasts. Given Al-bania’s reliance on exports, periodic validation and refinement of model-building are critical. The ARIMA findings recommend 3, 2, and 3, applied in forecasting. Forecast errors were tested, proving model efficiency. Performance may vary by situation, so model assessment over time is essential. These models help researchers forecast Albania's annual exports.