Is there any fast method to find the determinant of the following matrix

58 Views Asked by At

Is there any fast method (for manual computation by hand) to find the determinant of the matrix

$$[D \alpha ]_{i,j} = \begin{cases} 1 + (\frac{\partial f}{\partial x_i})^2 & ; i = j \\ \frac{\partial f}{\partial x_i} \cdot \frac{\partial f}{\partial x_j} & ; i \not = j \end{cases}, \text{where} \quad i = 1,...,n$$ and $f: \mathbb{R}^n \to \mathbb{R}^n$

1

There are 1 best solutions below

8
On BEST ANSWER

The matrix is equal to $$ I + \nabla f \cdot \nabla f^T, $$ where $\nabla f$ is a column vector, whose determinant equals $$ \det(I + \nabla f \cdot\nabla f^T) = \det( 1+ \nabla f^T \nabla f) = 1+ \nabla f^T \nabla f= 1 + \|\nabla f\|_2 ^2. $$ See Sylvester's determinant identity