Algorithmic data analysis of the nutritional impact on allergy during early life

Abstract

Nutrition during early life has been widely researched. A particular topic of interest is its relationship with allergy. Several features appear of impact, such as the weaning time point, the allergy state of the parents and the mode of delivery, while breastfeeding is considered the optimal nutritional option. A large clinical trial regarding early life nutrition and the impact on the health of infants was held in order to evaluate those hypotheses. Various parameters of the infants such as metabolites, gut microbial composition and immune system parameters were determined by the analysis of blood and stool samples. The outcome of this clinical trial is a high-dimensional dataset which was analysed in the context of this project by various machine learning algorithms. Tasks such as feature selection, classifcation and subgroup discovery contributed to reveal interesting properties of the infants which developed symptoms of Atopic Eczema/Dermatitis Syndrome and evaluate the effect that a nutritional intervention has during early life.