SHORT-TERM AND LONG-TERM PRODUCT DEMAND FORECASTING WITH TIME SERIES MODELS

This study explores the effectiveness of various time series models for short-term and long-term product demand forecasting, emphasizing the importance of accurate predictions in business operations. Demand forecasting is crucial for optimizing inventory levels, enhancing operational efficiency, and ensuring customer satisfaction. The paper categorizes forecasting into two primary types: short-term, which focuses on immediate operational needs, and long-term, which is essential for strategic planning and resource allocation. The analysis employs a rich dataset of historical sales data from a retail company, encompassing various influencing factors such as seasonal fluctuations and promotional impacts. The methodology includes data preprocessing steps to ensure data integrity, followed by the implementation of various time series models, including Moving Averages, Exponential Smoothing, ARIMA, Seasonal Decomposition of Time Series, Long-Term ARIMA, and SARIMA. The study also incorporates machine learning approaches to enhance forecasting accuracy. Evaluation metrics such as Mean Absolute Error, Mean Squared Error, Root Mean Squared Error, and Mean Absolute Percentage Error are utilized to assess model performance. The comparative analysis reveals that while traditional time series models are effective for short-term forecasting, advanced techniques like machine learning can significantly improve long-term predictions. The findings highlight the need for tailored modeling strategies based on specific business objectives and the importance of integrating external factors into forecasting models. Overall, this research contributes to the ongoing discourse in demand forecasting by identifying gaps in existing literature and suggesting areas for further exploration, such as the integration of short-term and long-term approaches and the incorporation of advanced techniques like AI and big data analytics.