Curve Fitting is the general method for using a line or curve to estimate the relationship between two associated numerical variables. The curve which best represents the relationship between the variables is called an approximating curve or a regression line.
Note: This Curve Fitting Tutorial is an introduction to various types of functions that can be used as approximating curves. See other Regression Tutorials for a more sophisticated approach including regression formulas, modeling assumptions, and diagnostics.
Typical Example of
Curve Fitting: Simple Linear Regression
fits a straight
line to the two numerical variables.
Other Examples: (The links provide a visual example.)
Quadratic Regression fits a parabola to the variables.
Cubic Regression fits a third degree polynomial.
Exponential Regression fits an exponential function (used for rapid growth or decay).
Sinusoidal
Regression fits a sine curve (used for periodic curves).
Specific Example: Assume that during a three-hour period spent outside, a person recorded the temperature and their water consumption. The points on the Regression Plot below represent the observations on 7 randomly selected summer days. The data appear to follow an approximate linear trend. An approximating curve is shown by the straight line on the scatterplot.
Return to Regression Level I Menu Dictionary