Classification is a supervised learning technique which requires prior knowledge of the groups involved. A dataset with correctly labelled or classified data is used as a starting block, then divided into a training set which enables analysts to verify the authenticity and of a model. Once done, the model can be used to classify new observations.