An AI-powered e-nose system using a density-based clustering method for identifying adulteration in specialty coffees

Abstract

Coffee is a valuable commodity with diverse varieties and unique specifications, particularly in Indonesia, a coffee-producing country. Ensured classification and quality of coffee throughout the supply chain are crucial for industry stakeholders, particularly in terms of flavor and aroma. This study focuses on the authenticity of coffee, particularly specialized coffee beans from microclimates that may be impacted by climate change. Advanced technologies, such as electronic nose (E-Nose) systems paired with machine learning techniques, play a significant role in achieving accurate and efficient authentication. Furthermore, the research addresses adulteration issues, such as blending different types of coffee beans, which can discourage farmers from participating in coffee cultivation. In this study, we develop an authentication method leveraging the olfactory characteristics of coffee aroma, emphasizing the unique interplay between microclimate conditions and flavor profiles. The clustering method of density-based spatial clustering of applications with noise (DBSCAN) technique is used to authenticate tropical climate specialty coffee and recognize adulterated specialty coffee beans. In our case study, adulteration in Arabica and Luwak specialty coffees was successfully detected using the DBSCAN technique combined with dimensionality reduction methods. This work has demonstrated a novel method for identifying specialty coffee authenticity and adulteration using the E-Nose system paired with machine learning techniques.