Legen Beverage Quality Assessment Using Multi-Sensor Data Fusion and Machine Learning Algorithms

Legen is a traditional drink valued for its health benefits and as a raw material for high-value processed foods. However, its quality declines over time due to natural fermentation. Quality assessment is typically based on organoleptic parameters such as color, aroma, and taste, while laboratory-scale methods remain costly and time-consuming. This study introduces a multi-sensor system—comprising a gas sensor array, a potentiometric array, and an RGB sensor—integrated with machine learning, including Neural Networks (NN) and Support Vector Machine (SVM), to classify legen quality. Experiments evaluated the system’s performance across different sensor configurations, data fusion strategies, and algorithms. The combination of all three sensor arrays, low-level data fusion, and a NN model achieved the highest accuracy, reaching 98.1% in classifying legen quality. These findings demonstrate the potential of multi-sensor machine learning systems as efficient and reliable tools for automated food quality assessment.