IoT-based Beverage Fraud Detection: A Theoretical Review

Abstract

This theoretical review explores the foundational theories, frameworks, and models relevant to the design of an Artificial Intelligence (Al) and Internet of Things (IoT)-based system for detecting counterfeit and expired carbonated beverages in the Nigerian market. The growing incidence of beverage fraud, including dilution, mislabeling, and expiration concealment, necessitates the adoption of advanced detection mechanisms. This study explores the use of Bayesian Linear Regression (BLR) to analyze caffeine and CO: concentrations within established theories from food safety, sensor analytics, and machine learning. By comparing and critiquing traditional and modern approaches, the review highlights the strengths of integrated lot and Machine Learning (ML) technologies for scalable, real-time quality monitoring. The key finding from the review is that integrating lot-enabled sensors, BLR, and ML within the Fake Beverages Detection Systems (FaBEDs) framework offers a scalable, real-time, and interpretable approach for detecting counterfeit and expired carbonated beverages, particularly suitable for low-resource settings like Nigeria. This work contributes a critical perspective on how theoretical models can inform practical implementations for safeguarding public health and ensuring beverage supply chain integrity in developing economies.