The learning problem consists of finding a hypothesis in a learning model which closely approximates the target function using a training set from the target function.

Examples:

1. example 1
2. example 2
3. linear model
4. nonlinear model