Using “Machine Learning” Techniques in Increasing the Efficiency of Sales Forecasting in Albania

This paper investigates the utilization of a Machine Learning (ML) approach with the objective of selecting an appropriate model for sales forecasting. Within this study, three ML algorithms are examined: Simple Linear Regression, Gradient Boosting Regression, and Random Forest Regression. A comparative analysis of these algorithms is conducted using two performance metrics: Accuracy Score and Max Error. The significance of sales forecasting cannot be overstated, as it plays a critical role across various industries. Therefore, the application of ML technology is essential to mitigate potential financial losses resulting from inaccurate demand assessments. A retail company based in Albania, which provided historical data as input for the model, is utilized as a case study. The Random Forest Model demonstrates exceptional performance, characterized by minimal deviations between predicted and actual values. The findings of this research endeavor present a pioneering initiative that holds significant potential for enhancing the forecasting of future sales and delivering substantial benefits to firms operating in the Albanian market.