| Minimizing squared error |
|
|
| vector of |
|
| target
values |
|
|
|
| optimal |
|
| weights |
|
|
|
|
|
| the
transposed |
|||
| design
matrix has |
|||
| one
input vector |
|||
| per column |
|||
|
|
|
|
| inverse
of the |
|||
| covariance |
|||
| matrix
of the |
|||
| input
vectors |
|||