Statistics Assignment
Tahmeed Muhtasim Nisorgo
2411022102
Physics
June 18, 2025
Normal Distribution:
Previously, we have been dealing with only discrete random variables. However,
there are also distributions of continuous random variables, given by their
probability density function. One such distribution is called the Normal Distribution
(also known as Gaussian Distribution)
Normal Distribution
We say a random Variable X has a normal Distribution if its pdf is
given by the formula
" 2 #
1 1 x−µ
f (x) = √ exp − , f or − ∞ < x < ∞
2π · σ 2 σ
where λ is a positive constant. Also, for any Normal random
variable, E(X)=λ
Figure 1: Normal Distribution
1
Mathematical Expectation of a Normal Distribution
The mathematical expectation for a Normal distribution is,
" #
Z ∞
1 1 xi − µ 2
k
E(X ) = xki √ exp − dx (1)
−∞ 2π · σ 2 σ
" #
1 Z ∞
k −y 2 xi − µ
=√ (µ + σy) exp σdy ;y = (2)
2π · σ −∞ 2 σ
" #
Z ∞
k 1 −y 2
= (µ + σy) √ exp dy (3)
−∞ 2π 2
" #
Z ∞
1
1 −y 2
E(X) = (µ + σy) √ exp dy (4)
−∞ 2π 2
" # " #
Z ∞
1 −y 2 Z ∞
1 −y 2
=µ √ exp dy + σ y √ exp dy (5)
−∞ 2π 2 −∞ 2π 2
= µ(1) + σ · (0) (6)
=µ (7)
The mathematical expectation of a Normal distribution is,
E(X) = µ
2
Raw moments of a Poisson Distribution
The 2nd Raw moment of Normal distribution is as follows,
" #
2
Z ∞
21 −y 2
E(X ) = (µ + σy) √ exp dy (8)
−∞ 2π 2
" #
Z ∞
2 2 2 1 −y 2
= (µ + 2 · µ · yσ + y σ ) √ exp dy (9)
−∞ 2π 2
" # " # " #
Z ∞
2 1 −y 2 Z ∞
1 −y 2 Z ∞
2 2 1 −y 2
= µ √ exp dy + µ · σy √ exp dy + σ y √ exp dy
−∞ 2π 2 −∞ 2π 2 −∞ 2π 2
(10)
" #
2
Z ∞
1 −y 2
2 2
= µ (1) + 2µσ(0) + σ y √ exp dy (11)
−∞ 2π 2
h 2 ilast integral we use integration by parts, considering, u = y and
Now on the
y √2π exp −y2 dy = dv
1
" #
−y 2
∴ v = −exp
2
du = dy
using substitution by parts,
du
Z Z Z
= u · dv = uv − dv dx
∞
dx
Z ∞
1 −y 2 −y 2
=√ −ye 2 − e 2 dy
2π −∞ −∞
The first sum is just 0 and last integral is 1. Putting the respective values
in (11) we get,
E(X 2 ) = µ2 (1) + 0 + σ 2 (1)
E(X 2 ) = µ2 + σ 2
Similarly, The 3rd and 4th Raw moment can be determined,
E(X 4 ) = 3σ 2 µ + µ3
E(X 4 ) = 3σ 4 + 6σ 2 µ2 + µ4
E(X) = µ
E(X 2 ) = µ2 + σ 2
E(X 3 ) = 3σ 2 µ + µ3
E(X 4 ) = 3σ 4 + 6σ 2 µ2 + µ4
3
Moment Generating Function of Normal Distribution
Z ∞
1 −(x−µ)2
Mx (t) = ext √ e 2σ2
−∞ 2π · σ
Z ∞
1 −y 2 x−µ
= e(µ+σy)t √ e 2 σdy ∴ =y
−∞ 2π · σ σ
Z ∞
1 −y 2
= eµt eσyt √ e 2 dy
−∞ 2π
Z ∞
1 −y 2 −σ 2 t2 σ 2 t2
= eµt eσyt √ e 2 e 2 e 2 dy
−∞ 2π
Z ∞
2
σ t 2 1 −(y−σt)2
= eµt e 2 √ e 2 dy
−∞ 2π
" #
2
2µt+σ t 2
2µt + σ 2 t2
=e 2 = exp
2
1st, 2nd, 3rd and 4th Raw moment of Normal Distribution
∂M ∂ 2µt+σ2 t2
= (e 2 )
∂t t=0
∂t
2µt+σ 2 t2
=e 2 · (µ + σ 2 t)
=µ
∂ 2M ∂ 2µt+σ2 t2
= (e 2 · (µ + σ 2 t))
∂t2 t=0
∂t
2µt+σ 2 t2 2µt+σ 2 t2
= (µ + σ 2 t)2 e 2 + σ2 · e 2
2µt+σ 2 t2
= (µ2 + 2µσ 2 t + σ 4 t2 + σ 2 )e 2
= µ2 + σ 2
∂ 3M ∂ 2 2µt+σ 2 t2
= (µ + 2µσ 2 t + σ 4 t2 + σ 2 )e 2
∂t3 t=0
∂t
2µt+σ 2 t2 2µt+σ 2 t2
= (2µσ 2 + 2σ 4 t)e 2 + (µ2 + 2σ 2 t + σ 4 t2 + σ 2 )(µ + σ 2 t)e 2
= 3σ 2 µ + µ3
∂ 4M ∂ 2µt+σ 2 t2 2µt+σ 2 t2
= (2µσ 2 + 2σ 4 t)e 2 + (µ2 + 2σ 2 t + σ 4 t2 + σ 2 )(µ + σ 2 t)e 2
∂t4 t=0
∂t
2µt+σ 2 t2 2µt+σ 2 t2
= (2σ 4 )e 2 + (2µσ 2 + 2σ 4 t)(µ + σ 2 t)e 2 +
2µt+σ 2 t2
(µ3 + µσ 4 t2 + 2µ2 σ 2 t + µσ 2 + µ2 σ 2 t + σ 6 t4 + 2µσ 4 t2 + σ 4 t)(µ + σ 2 t)e 2 +
2µt+σ 2 t2
(2µσ 4 t + 2µ2 σ 2 + µ2 σ 2 + 4σ 6 t3 + 2σ 4 t + σ 4 )e 2
= 3σ 4 + 6σ 2 µ2 + µ4
In this way we can generate all the raw moments of a distribution.
4
Central moments of Normal Distribution
Doing some mathematical jargon we can find the central moments of Normal Distribution
µ2 = µ′2 − µ′2
1
= µ2 + σ 2 − µ2
= σ2 [The 2nd cetnral moment is the Variance as expected]
µ3 = µ′3 − 3µ′2 µ′1 + 2µ′3
1
= 3σ µ + µ − 3(µ + σ 2 )(µ) + 2(µ)3
2 3 2
=0
µ4 = µ′4 − 4µ′3 µ′1 + 6µ′2 µ′2 ′4
1 − 3µ1
= 3σ 4 + 6σ 2 µ2 + µ4 − 4(3σ 2 µ + µ3 )(µ) + 6(µ2 + σ 2 )(µ)2 − 3(µ)4
= 3σ 4
Cumulant Generating Function
Cumulant generating function is described the function,
Kx (x) = loge (Mx (t)) = lnMx (t)
dr Kx (t)
κr =
dtr t=0
Cumulant generating function of Normal Distribution
Kx (x) = loge (Mx (t)) = lnMx (t)
2µt+σ 2 t2
= ln(e 2 )
2 2
σ t
= µt +
2
5
κr for r = 1, 2, 3, 4 for Normal Distribution
2 2
d(µt + σ 2t )
κ1 = =µ
dt t=0
2 2
d2 (µt + σ 2t )
κ2 = = σ2
dt2 t=0
2 2
d3 (µt + σ 2t )
κ3 = =0
dt3 t=0
2 2
d4 (µt + σ 2t )
κ4 = = 0 = 3σ 4 − 3(σ 2 )2
dt4 t=0
Using cumulant generating functions we can determine the central moments of
any distribution.
Skewness and Kurtosis of Normal Distribution
We know,
µ23
β1 =
µ32
µ4
β2 = 2
µ2
v
u 2
u µ3 q
γ1 = t = β1
µ32
γ2 = β2 − 3
Putting the respective values we can determine the skewness and kurtosis of
a Normal distribution
(0)
β1 = =0
(σ 2 )3
0
β2 = 2 2 = 0
(σ )
q
γ1 = β1 = 0
γ2 = β2 − 3 = −3