Data Mining Dynamic Hybrid Model for Logistic Supplying Chain: Assortment Setting in Fast Fashion Retail

Data science tools have been used in many fields as effective techniques for data analysis. Artificial intelligence, machine learning and data mining made the buzz on both industrial and scientific communities, pushing the researchers to look for the potential value added they might have by using the tool. Fast fashion retailers joined the vague too, but still have many untapped fields to work on. In this paper, we work on the assortment problem for a worldwide fast fashion retailer, who sells a large quantity of products, with a wide range of models and different regions of sales. Every region has its own features in term of habits, clothing choices and trends, thus the retailer uncertainty to dispatch its inventory in an optimal way, as to meet the expectations of the customers, building the consumers loyalty and maximize the sales. The proposed procedure is programmed with Python and orange software, and then tested with data instances, inspired from real cases.