IMPROVEMENT OF DEMAND FORECASTING ACCURACY USING MACHINE LEARNING MODEL: (A CASE OF OROMIA SEED ENTERPRISE)

Abstract

The abstract discusses a study aimed at improving demand forecasting accuracy for Oromia Seed Enterprise by leveraging machine learning models. The enterprise faces challenges in business operations due to the uncertainty of future conditions, making accurate demand forecasting essential for better decision-making and operational optimization. The study analyzes forecasting variables, rates their impacts, and uses machine learning techniques support vector regression (SVR) and random forest regression (RFR) to enhance forecasting performance. Data analysis and modeling were conducted using Python within the Anaconda environment.The SVR model achieved a perfect R-squared score of 1.0, indicating excellent accuracy, while the RFR model demonstrated a score of 0.989. Various error metrics, such as mean absolute error and root mean squared error, were employed to evaluate the models. SVR showed lower error values compared to RFR, emphasizing its superior predictive performance. The findings suggest that machine learning models, particularly SVR, can significantly improve demand forecasting accuracy.The research highlights the importance of addressing variables influencing forecasting, with respondents assigning moderate impact ratings to these factors. The study concludes that enhancing demand forecasting through machine learning could significantly boost operational efficiency in seed enterprise operations. It recommends that businesses adopt such models to optimize productivity and decision-making across various aspects of the enterprise.