Regression planar
0 likes562 views
Planar regression calculates the best fit plane through 3 or more data points by minimizing the residuals between the plane and points using least squares minimization. The equation for the plane is determined by taking the derivative of the residuals equation and setting it equal to zero, yielding a system of equations that can be expressed in matrix form and solved for the plane coefficients.
1 of 2
Download to read offline
![Planar/Bilinear Least Squares Regression
Planar regression calculates the best fit plane through a group of 3 or more data points. The plane is calculated by
minimizing the residuals (or errors) between the plane and the original points using least squares minimization.
The least squares minimization equation is:
∑( ( ) )
Where are the observed values and ( ) is the y-value of the surface at . The equation of the plane is:
Plugging this value in to the regression equation gives
∑( )
To find the minimum residual error, the derivative of the residuals equation must be zero, which means all of the
partial derivatives with respect to each coefficient must be equal to zero.
∑( )
∑( )
∑( )
These equations can be expressed in matrix form:
[
∑ ∑ ∑
∑ ∑ ∑
∑ ∑ ∑ ]
[ ]
[
∑
∑
∑ ]
Solving for and :](https://image.slidesharecdn.com/regressionplanar-180303041916/85/Regression-planar-1-320.jpg)
![[ ]
[
∑ ∑ ∑
∑ ∑ ∑
∑ ∑ ∑ ] [
∑
∑
∑ ]
Property of ahinson.com – Last Updated December 9, 2011](https://image.slidesharecdn.com/regressionplanar-180303041916/85/Regression-planar-2-320.jpg)
More Related Content
Recently uploaded (20)
Ad
Ad
Regression planar
- 1. Planar/Bilinear Least Squares Regression Planar regression calculates the best fit plane through a group of 3 or more data points. The plane is calculated by minimizing the residuals (or errors) between the plane and the original points using least squares minimization. The least squares minimization equation is: ∑( ( ) ) Where are the observed values and ( ) is the y-value of the surface at . The equation of the plane is: Plugging this value in to the regression equation gives ∑( ) To find the minimum residual error, the derivative of the residuals equation must be zero, which means all of the partial derivatives with respect to each coefficient must be equal to zero. ∑( ) ∑( ) ∑( ) These equations can be expressed in matrix form: [ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ] [ ] [ ∑ ∑ ∑ ] Solving for and :
- 2. [ ] [ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ] [ ∑ ∑ ∑ ] Property of ahinson.com – Last Updated December 9, 2011