turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Blogs
- :
- JMPer Cable
- :
- Understanding cubic splines

Article Options

- Subscribe to RSS Feed
- Mark as New
- Mark as Read
- Bookmark
- Subscribe
- Email to a Friend
- Printer Friendly Page
- Report Inappropriate Content

May 24, 2017 6:22 AM
(837 views)

Understanding cubic splines

A cubic spline is a smooth, continuous function that represents the relationship between two numeric variables. It consists of a series of third-order polynomials fit together in a piecewise fashion, under the condition that the function is continuous and differentiable at the knots, or connection points.

JMP offers spline fits in the Fit Y by X platform. For example, using the Fitness.jmp sample data table, I can fit a cubic spline for the Oxy column by the RunPulse column.

You can get the coefficients of the polynomial functions by clicking the inverted red arrow hot-spot for “Smoothing Spline Fit” and selecting “Save Coefficients”. A new data table is generated containing an X column that gives the “RunPulse” values (in this case) that are taken as the knots where the individual cubic polynomial functions are connected. The remaining columns in the data table are the coefficients for each cubic spline: “A” is the intercept term, “B” is the coefficient for the linear term, “C” is the coefficient for the quadratic term, and “D” is the term for the cubic term for the variable (X-O) where O is the value of the knot at the beginning of the interval:

To better understand the concept of the spline fit, I updated the coefficients table to include “Oxy” values using the "Tables->Update" option:

Then I recreated the spline fit using the coefficients, extending the polynomials beyond the knot points:

The points and dashed vertical lines show the knot points, and the individual polynomial curves are represented by different colors. Finally, you can see the spline fit itself when the polynomials are truncated at the knots:

Here, you can see each segment of the final cubic spline, and how they are connected together. Hopefully, this explanation will take any mystery out of cubic splines and help to understand this smoothing fit.

Article Tags

You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.