Regression can be used to smooth data by fitting the data to a function.
It is a statistical measure that attempts to determine the strength of the relationship between one dependent variable (usually denoted by Y) and a series of other changing variables (known as independent variables).
It is a form of mathematical model that reflects results in a curve between two variables (X & Y) rather than a straight-line relationship (as in the case of a linear regression, which are usually simpler with expression such as Y= a + bx).
The main characteristic of a non-linear regression is that the prediction equation does not linearly depend on the unknown parameters.
This type of regression uses functions such as trigonometry, logarithm and exponential.