Could anyone help me with this result?
To solve the equation $P(x)=\theta$, we can aply a modified Newton method such that the itarations are given by the formula $$x_{n+1}=x_n-[P'(x_1)]^{-1}P(x_n) \quad (1)$$ supposing that $P$ is Fréchet differentiable on a convex region $D$, and for some $x_1 \in D$, the inverse $[P'(x_1)]^{-1}$ exists.
Suppose that $x_2$ computed with $(1)$ belongs to $D$, also suppose that $$\rho=\lVert [P'(x_1)]^{-1}\rVert \cdot \sup_{x\in D}\lVert P'(x_1)-P'(x)\rVert<1$$ and the ball $$S=\{x:\lVert x-x_2\rVert<\frac{\rho}{1-\rho}\lVert x_1-x_2\rVert\}$$ is a subset of $D$. Show that this modified method converges to a solution $x_0\in S$.
Any help would be appreciated.