Leveraging Graph Neural Networks with procurement data

Graph Neural Networks (GNNs) are particularly effective when dealing with non-Euclidean data representations without losing their inherent meaning. In our scenario, we have a network of connections among over 10,000 suppliers and aim to generate accurate recommendations for a new supplier or address inquiries such as, What is the best alternative supplier to switch to? and What are the characteristics and motivations for acquiring a supplier? We will delve into the theoretical aspects of GNNs and present the findings obtained.