The primary objective of supervised learning is to generalize a model from labeled training data. Once a model has been trained, it allows users to make predictions about unseen future data. Here, by labeled training data, we mean the training examples know the associated output labels. Hence, it is referred to as supervised learning. The learning process can be thought of as a teacher supervising the entire process. In such a learning process, we know the correct answer initially, and the students learn enough iteratively over time and try to answer unseen questions. The errors in the answers are corrected by the teacher. The process of learning stops when we can ensure the performance of the student has reached an acceptable level. 

In supervised learning, we have input variables (x_i) and output variables (Y_i). With this, we can learn a function, f, as shown by the following equation:

The objective is to learn a general mapping function, f, so that the function can predict the output variable, Y, for any new input data, x. Supervised learning algorithms can be categorized into two groups, as follows:

• Regression
• Classification

Let's take a brief look at these.