Machine Learning Performance on IoT Device

The Internet of Things (IoT) is transforming device connectivity and interaction. Integrating machine learning (ML) further enhances IoT capabilities, enabling smarter, more adaptive systems. However, deploying ML on resource-constrained edge devices presents significant challenges. This paper evaluates the performance of several ML classifiers on a Raspberry Pi 3 (RPi3), focusing on processing time, memory usage, and energy consumption. Our results demonstrate the feasibility of deploying ML on the RPi3, revealing that computational and energy costs are primarily determined by the chosen classification algorithm and, secondarily, by dataset size. Random Forest and Decision Tree models emerged as the most efficient in terms of energy consumption and processing time, while also achieving high accuracy.