the samples are described by a vector of features

we need to split the data into a training set and a test set

the data in the training set must be labeled

we need to evaluate the results using various metrics