Technological advances have significantly increased the amount of data available in many fields.  In doing so, this has created the infamous n<p problem, where the sample size, n, is much smaller than the number of parameters, p.  Classical methods for classification can no longer be implemented and several new methods have evolved.  Methods such neural networks, support vector machines, and random forests have revolutionized the supervised learning problem of classification.  In this talk, we will focus on the motivation behind the Random Forest algorithm, and illustrate this algorithm on a metabolomics data set.