The goal in classification is to assign a class label to each element of a data set. Of course, we would want this done with the highest accuracy possible. For us, there are only two classes or labels: positive sentiment and negative sentiment. At a high level, the process to build a classifier (and many other machine learning models) is this: