- #1
Trifis
- 167
- 1
I was wondering if such an approximation is possible and plausible...
The first term would have to look sth like this: [itex]\vec{f}[/itex]([itex]\vec{x_{0}}[/itex]) + [itex]\textbf{J}[/itex][itex]_{[itex]\vec{f}[/itex]}[/itex]([itex]\vec{x_{0}}[/itex])[itex]\cdot[/itex]([itex]\vec{x}[/itex]-[itex]\vec{x_{0}}[/itex])
No clue about the second term though...
We would have to calculate the Jacobian of the Jacobian (like we calculate the Jacobian of the Gradient to get the Hessian for the second term of the regular case of scalar fields f: ℝ[itex]^{n}[/itex]→ℝ) or sth...
The first term would have to look sth like this: [itex]\vec{f}[/itex]([itex]\vec{x_{0}}[/itex]) + [itex]\textbf{J}[/itex][itex]_{[itex]\vec{f}[/itex]}[/itex]([itex]\vec{x_{0}}[/itex])[itex]\cdot[/itex]([itex]\vec{x}[/itex]-[itex]\vec{x_{0}}[/itex])
No clue about the second term though...
We would have to calculate the Jacobian of the Jacobian (like we calculate the Jacobian of the Gradient to get the Hessian for the second term of the regular case of scalar fields f: ℝ[itex]^{n}[/itex]→ℝ) or sth...
Last edited: