inequality of the chi squared distribution function with d degree of freedom and the one with two

17 Views Asked by At

Good morning, I can't find how to prove this inequality :

$F_d(x) ≥ (⌊d/2⌋ + 1)*F_2(x/(⌊d/2⌋ + 1))$

where $F_d(x) = P(\lVert Z_s \rVert^2 \leq x)$ with $Z_s \sim\mathcal{N}(0, I_d)$

I have the intuition we should group together two by two the component of the vector but then I don't know how to show it. I guess we have to use the fact that $F_d(x)$ is the distribution function of the $\chi^2(d)$ distribution.