Specification and Identification of Relationships between Products in the Food Segment

Abstract

There are numerous relationships between products in retail. Understanding these relationships is particularly valuable for optimizing e-commerce, master data management, and marketing, among other applications. However, these relationships are often not explicitly known. In this work, we address the problem of automatically detecting a wide range of product relationships within the food segment. To achieve this, we provide a detailed specification of relationships specific to food products. Additionally, we describe multi-class classification models for automatically determining these relationships.