Machine Learning Assisted Sanitizer Quality Identification Using Permittivity Measurement

Sanitizers play a crucial role in hygiene and the prevention of infectious diseases by reducing harmful microorganisms on surfaces. This paper focuses on alcohol-based sanitizers, specifically those containing ethanol and isopropyl alcohol, known for their rapid action and broad spectrum efficacy. The effectiveness of these sanitizers depends on the alcohol content, with water enhancing their performance. Traditional quality detection methods are accurate but time-consuming. This work proposes a support vector machine based classification model combined with a complex permittivity measurement system to assess sanitizer quality. The model achieves 97.2% accuracy for a frequency range of 0.2–20 GHz at 25°C, with improved accuracy of 99% in the 2–5 GHz range, identifying this as the optimal frequency range for sanitizer classification. This method is compatible with any permittivity measurement system with measurement error ±15%. Significant time savings in quality assessment is also achieved.