A comparative study of time series forecasting: traditional approaches vs. machine learning methods

This study aims to compares traditional time series forecasting methods with machine learning techniques using Super Store sales data from 2011 to 2014, focusing on the technology category and its accessories subcategory, which are organized into 52-week intervals over four years. With a total of 208 weeks, the study seeks to forecast sales for the next 4 weeks. The study aims to provide insights into the relative strengths and limitations of traditional methods in comparison to machine learning methods. Performance metrics such as Root Mean Squared Error (RMSE) and Mean Absolute Deviation (MAD) will be used to evaluate forecasting accuracy. By evaluating these metrics, this study hopes to help retail businesses improve forecasting methodologies for greater predictive accuracy and more informed operational strategies.