I expect "binomial (4, .5)" refers to the binomial distribution with
n=4, p = 0.5 = q
Hence mean μ = np = 2
and standard deviation σ = npq = 1
Now Chebyshev's inequality is:
"Let X be a random variable with expected value μ and variance σ². Then for any real number k ≥ 1
P(| x- ( μ )| <kσ) ≥ 1 - (1/k²)"
which in this case translates to
P(|X - 2| < 2*1) ≥ 1 - 1/4 = 3/4
Now the event |X-2|≥2 is the complement of |X - 2| < 2
Hence
P(|X-2|≥2) = 1 - P(|X - 2| < 2)
Now since P(|X - 2| < 2) ≥ 3/4
therefore -P(|X - 2| < 2) ≤ -3/4
and so 1-P(|X - 2| < 2) ≤ 1-3/4
i.e. P(|X-2|≥2) ≤ 1/4, or 0.25
How does this compare with exact probability? i.e. the probability that X = 0 or 4
[ |X-2| ≥ 2 means either X-2 ≥ 2, so X ≥ 4; or X-2 ≤ -2, so X ≤ 0. Of the possible values of X (0, 1, 2, 3, 4), the only ones that satisfy this are 0 and 4]
The probabilities for the binomial distribution are given by the terms in the expansion of (p+q)^n,
hence we look at the expansion of (0.5 + 0.5)⁴, and so
The full list of probabilities for this distribution is
Answers & Comments
Verified answer
I expect "binomial (4, .5)" refers to the binomial distribution with
n=4, p = 0.5 = q
Hence mean μ = np = 2
and standard deviation σ = npq = 1
Now Chebyshev's inequality is:
"Let X be a random variable with expected value μ and variance σ². Then for any real number k ≥ 1
P(| x- ( μ )| <kσ) ≥ 1 - (1/k²)"
which in this case translates to
P(|X - 2| < 2*1) ≥ 1 - 1/4 = 3/4
Now the event |X-2|≥2 is the complement of |X - 2| < 2
Hence
P(|X-2|≥2) = 1 - P(|X - 2| < 2)
Now since P(|X - 2| < 2) ≥ 3/4
therefore -P(|X - 2| < 2) ≤ -3/4
and so 1-P(|X - 2| < 2) ≤ 1-3/4
i.e. P(|X-2|≥2) ≤ 1/4, or 0.25
How does this compare with exact probability? i.e. the probability that X = 0 or 4
[ |X-2| ≥ 2 means either X-2 ≥ 2, so X ≥ 4; or X-2 ≤ -2, so X ≤ 0. Of the possible values of X (0, 1, 2, 3, 4), the only ones that satisfy this are 0 and 4]
The probabilities for the binomial distribution are given by the terms in the expansion of (p+q)^n,
hence we look at the expansion of (0.5 + 0.5)⁴, and so
The full list of probabilities for this distribution is
(0.5)⁴ + 4*(0.5)³(0.5) + 6*(0.5)²(0.5)² + 4*(0.5)(0.5)³ + (0.5)⁴
P(X=0) = (0.5)⁴
Also P(X=4) = (0.5)⁴
Therefore
P(X=0) + P(X=4) = 0.125
which is less than 0.25 and so agrees with Chebyshev's inequality.