next up previous contents
Next: 3.2.3 Numerical Aspects Up: 3.2 Response Surface Methodology Previous: 3.2.1 Mathematical Background


3.2.2 Solving the Least-Squares Problem

The response surface must minimize the least-squares of $\vec{\epsilon}$

\begin{displaymath}
\mathop{\rm min}\limits \vec{\epsilon}^{\cal T} \vec{\epsilon}
.
\end{displaymath} (3.7)

Calculation of the first derivatives3.3 gives

$\displaystyle \frac{\partial \vec{\epsilon}^{\cal T} \vec{\epsilon}}
{\partial \vec{a}}$ = $\displaystyle \frac{\partial} {\partial \vec{a}}
\left (-\mathcal{Z}\vec{a} + \vec{y}\right )^{\cal T}
\left (-\mathcal{Z}\vec{a} + \vec{y}\right ) =$ (3.8)
  = $\displaystyle \frac{\partial} {\partial \vec{a}}
\left (\vec{a}^{\cal T} \mathc...
...vec{a}^{\cal T}\mathcal{Z}^{\cal T}\vec{y} +
\vec{y}^{\cal T}\vec{y} \right ) =$ (3.9)
  = $\displaystyle \left ( \mathcal{Z}^{\cal T} \mathcal{Z} +
\left ( \mathcal{Z}^{\...
... (\vec{y}^{\cal T} \mathcal{Z} \right )^{\cal T} -
\mathcal{Z}^{\cal T}\vec{y}=$ (3.10)
  = $\displaystyle 2 \mathcal{Z}^{\cal T}\mathcal{Z}\vec{a} -
2 \mathcal{Z}^{\cal T} \vec{y}$ (3.11)

Setting (3.11) to zero gives the formula for calculating the response surface

\begin{displaymath}
\vec{a} = \left ( \mathcal{Z}^{\cal T} \mathcal{Z} \right )^{-1}
\mathcal{Z}^{\cal T} \vec{y}
.
\end{displaymath} (3.12)

The matrix $\left ( \mathcal{Z}^{\cal T} \mathcal{Z} \right
)^{-1} \mathcal{Z}^{\cal T}$ is also called the left-pseudo-inverse. It is a left-inverse of $\mathcal{Z}$ because $\mathcal{Z}$ has to be multiplied by this matrix from the right side to yield the identity matrix $\mathcal{I}$

\begin{displaymath}
\left ( \mathcal{Z}^{\cal T} \mathcal{Z} \right
)^{-1} \mathcal{Z}^{\cal T} \ \mathcal{Z} = \mathcal{I}
.
\end{displaymath} (3.13)

For a minimum in least-squares of the error $\vec{\epsilon}$ the second derivative

\begin{displaymath}
\frac{\partial^2 \vec{\epsilon}^{\cal T} \vec{\epsilon}}
{\...
...\cal T} \partial \vec{a}} =
\mathcal{Z}^{\cal T} \mathcal{Z}
\end{displaymath} (3.14)

must be nonsingular. This is the case if the if the matrix $\mathcal{Z}$ has full column rank. For practical use the vector $\vec{a}$ can be calculated by a Gauß-solver with the system matrix $\mathcal{Z}^{\cal T} \mathcal{Z} $ and the data vector $\mathcal{Z}^{\cal T} \vec{y}$.



Footnotes

... derivatives3.3
Some rules for the derivatives of scalar functions with respect to a vector used in (3.10) are given in Appendix B.4.

next up previous contents
Next: 3.2.3 Numerical Aspects Up: 3.2 Response Surface Methodology Previous: 3.2.1 Mathematical Background

R. Plasun