Store Clustering in Supply Chains using ART Network

Abstract

When companies tend to manage the information that flows through a product supply chain, it becomes necessary to work closely with suppliers, logistics providers, distributors, and retailers to collect and manage information about customer demand, sales orders, distribution schedules, production planning, manufacturing, sourcing, and product. Getting this right can make the difference between success and failure, profit and loss, growth and loss of market share. Managing such a complex supply chain requires sophisticated data analysis that can allow the retailer to anticipate demand, and particularly shifts in that demand, ahead of time. Hence clustering data in a more appropriate manner would facilitate a very effective data analysis and further data mining. This paper proposes a suitable clustering technique namely the ART (Adaptive Resonance Theory) network to classify supply chain data under various identified analytic subject areas. Though there is a majority of clustering algorithms, a neural network is suitable for both deep knowledge discovery and complex predictive model construction.