An Improved Multi Classification of Milk Quality Using Machine Learning

Improving milk quality and safety through advanced prediction and monitoring systems addresses critical public health and sustainability challenges, aligning closely with Sustainable Development Goals - SDG 3 (Good Health and Well-being) and SDG 12 (Responsible Consumption and Production). Traditional milk quality testing methods, though effective, are often labor-intensive and slow, limiting scalability. Integrating machine learning technologies facilitates real-time, high-accuracy assessments by analyzing attributes of milk, enabling more efficient detection of adulteration and contamination. In this study, first analyzed various machine learning algorithms and identified that the Decision Tree algorithm offers higher accuracy than other ML models. The proposed system enhances milk quality classification using a combination of Principal Component Analysis (PCA) and the Decision Tree algorithm by applying Robust Scaling technique for handling data outliers. Parameters such as pH, temperature, taste, odor, fat content, turbidity, and color are used to classify milk into high, medium, and low quality. Results indicate that the Decision Tree with PCA provides improved accuracy for milk quality analysis.