A Machine Learning Approach to Forecasting Demand in Fashion Industry

Fashion apparel retail is lucrative, dynamic, and unpredictable, as it is influenced by a wide range of factors, including culture, economics, social media, and fashion trends. Due to these influences, coupled with the industry's inherent need to innovate, track customer preferences and global trends, and introduce new products, demand and sales forecasting are particularly challenging. Since the fashion apparel industry has a high degree of uncertainty, the methods available for demand forecasting in other industries are more difficult to apply. Data analysis and forecasting are becoming increasingly automated through the use of ML (Machine Learning) tools and algorithms. Our research suggests such tools and algorithms might be applied to fashion. This proposed model presents a forecasting method that can be used to develop a collection and manage production and supply chains by fashion retailers, based on their existing products as well as the model's assessment.