Random Forest machine learning method
We used Random Forest (RF) for data integration and individual data set analysis. RF is a machine learning ensemble method in conjunction with multiple learning algorithms to obtain better predictive performance18. RF can be used for both classification and regression. In our analysis we used RF for classification using the feeding method (exclusively breastfed vs exclusively formula-fed) as outcome variable and treating each of the data sets separately. We used ntree = 500 and mtry = square root of variables in our models. We used two packages for RF analysis (randomForest and varSelRF) in R (v3.6.1).