data mining task predictive modeling task Learning a (Probabilistic) Predictive Model. In this task, we are given a dataset that consists of examples of the form (d, c), where each d is of type Td and each c is of type Tc. We will refer to d as the description and c as the class or target. To learn a predictive model means to find a mapping from the description to the target, m :: Td → Tc, that fits the data closely. This means that the observed target values and the target values predicted by the model, i.e., c and ˆc = m(d), have to match closely.