Variation comparison between the F-distribution and the normal distribution

Abstract: Let

X_{d_{1}, d_{2}}

be an

F

-random variable with numerator and denominator degrees of freedom

d_{1}

and

d_{2}

, respectively. We investigate the inequality:

P | X_{d_{1}, d_{2}} - E [X_{d_{1}, d_{2}}] | l e s q r t m V a r (X_{d_{1}, d_{2}}) g e P | W - E [W] | l e s q r t m V a r (W)

, where

W

is a standard normal random variable or a

c h i^{2} (d_{1})

random variable. We prove that this inequality holds for

d_{1} i n 1, 2, 3, 4

and

5 l e d_{2} i n m a t h b b N

.

This page was built for publication: Variation comparison between the $F$-distribution and the normal distribution