Research on the Classification of Electromagnetic Radiation Safety Levels of Mobile Communication Terminals Using Unsupervised Learning Models

Specific Absorption Rate (SAR) is a critical metric for assessing the electromagnetic radiation safety of devices, and it plays a significant role in international standards and certifications. Leading organizations, such as the International Commission on Non-Ionizing Radiation Protection (ICNIRP), the Institute of Electrical and Electronics Engineers (IEEE), and several countries have established mandatory standards and clearly defined SAR limits. However, there is a lack of a more detailed grading system, especially concerning highly sensitive groups such as infants, pregnant women, and others, whose sensitivity to electromagnetic radiation and potential health risks are much higher than that of the general population. This study, based on 1,119 SAR measurement data of mobile communication terminals published by the French National Frequency Agency (ANFR) from 2012 to 2024, applies unsupervised learning methods to uncover the natural distribution characteristics of SAR data. By integrating health protection standards from multiple countries, clinical research findings, and certification practice, this study proposes a four-level SAR safety classification proposal, aiming to provide a foundation for the development of a labeling and certification system for the electromagnetic radiation safety levels of mobile communication terminals.