In a supervised scenario, the task of the model is to find the correct label of a sample, assuming that the presence of a training set is correctly labeled, along with the possibility of comparing the estimated value with the correct one. The term supervised is derived from the idea of an external teaching agent that provides precise and immediate feedback after each prediction. The model can use such feedback as a measure of the error and, consequently perform the corrections needed to reduce it.
More formally, if we assume a data generating process, the dataset is obtained as:
As discussed in the previous ...