Analyzing Milk Foam Using Machine Learning for Diverse Applications

In the beverages industry, milk foaming is done to enhance the flavor, texture, and visual appeal of milk-based beverages. It is very crucial to study milk foam properties not just to create visually appealing and rich in taste beverages but also to estimate the adulterants present in it. Machine learning is being used in every field nowadays as it can analyze large datasets quickly and help in making data-driven decisions. This paper is a demonstration of how a futuristic apparatus will detect the best type of milk for beverages and identify milk adulteration using machine learning. In the current study, machine learning methods are employed to assess milk foam properties. This study aims to choose the best type of milk for foam-based milk beverages preparations and detect surfactants often used in low concentrations for foaming but act as adulterants at high concentrations. Surfactants alter the foaming properties of milk in different ways depending on their charge and are therefore used in the dairy industry. By using machine learning techniques, the impact of three different surfactants, having distinct ionic properties, on three distinct types of milk have been analyzed. It was found that foaming properties of milk were highly correlated to each other. "Random forest classifier" turned out to be the most effective among all the machine learning models in both the tasks. Heating and addition of sodium dodecyl sulfate (SDS) improved foaming. The findings of this study can be used for deriving valuable insights about the dairy industry.