Predicting Unreported Micronutrients From Food Labels: Machine Learning Approach (Preprint)

Abstract

BACKGROUND Micronutrient deficiencies are a global health issue, with over two billion people suffering from deficiencies in vitamins and minerals. Food labels can help improve diets by providing information about the nutritional content of foods to consumers. However, food labels often only include a limited amount of information due to size and readability constraints and may not provide information about all micronutrients. OBJECTIVE This paper aims to examine the possibility of using machine learning algorithms to predict unreported micronutrients from the existing food label data. If unreported micronutrients can be accurately predicted from existing food labels using predictive models, such models can be integrated into mobile applications to provide consumers with additional micronutrient information about foods and help them make more informed diet decisions. METHODS Nutrition data from a total of 5,624 foods are used to train a wide array of machine learning algorithms to predict unreported vitamins and minerals from existing food label data. Models are evaluated using repeated cross-validations to ensure that they are not overfitting. RESULTS According to the results, while predicting the exact quantity of vitamins and minerals is challenging with regression R-squared varied in a wide range from 0.28 (for magnesium) to 0.92 (for manganese), the classification models could accurately predict the category (“low,” “medium,” or “high”) level of all minerals and vitamins with accuracies exceeding 0.80. The highest classification accuracies for specific micronutrients are achieved for vitamin B12 (0.94) and phosphorus (0.94), while the lowest are for vitamin E (0.81) and selenium (0.83). CONCLUSIONS The feasibility of predicting unreported micronutrients from existing food labels is demonstrated, for the first time, in this study. The viability of this approach is shown to have the potential to significantly improve consumer knowledge about the micronutrient content of the foods they consume. When integrated into mobile applications, this capability can be made more accessible and engaging for consumers. The implications of these findings for public health are significant, highlighting the potential of technology to enhance consumers’ understanding of the micronutrient content of their diet.