Math 2225 (MHU) 1

Lecture-11-14
Binomial Distribution
#Question 1: Define binomial distribution. Prove that, total density function of binomial distribution is unity.
Binomial Distribution: If 𝑥 is a random variable having density function
𝑛
( ) 𝑝 𝑥 𝑞 𝑛−𝑥 ; 𝑥 = 0, 1, 2, ⋯ ⋯ 𝑛, 𝑝 + 𝑞 = 1.
𝑃(𝑥) = { 𝑥
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Then 𝑥 is called binomial distribution and its density function is called probability density function.
Where 𝑝is the probability of success of an event in a trail and 𝑞is the probability of its failure, 𝑛 is
the number of independent trials.
𝑛
Proof: We know, 𝑃(𝑥) = ( ) 𝑝 𝑥 𝑞 𝑛−𝑥 ; 𝑥 = 0, 1, 2, ⋯ ⋯ 𝑛, 𝑝 + 𝑞 = 1.
𝑥
𝑛
∴ ∑𝑥=0 𝑃(𝑥) = ∑𝑛𝑥=0 ( ) 𝑝 𝑥 𝑞 𝑛−𝑥
𝑛
𝑥
𝑛 0 𝑛−0 𝑛 𝑛 𝑛
= ( )𝑝 𝑞 + ( ) 𝑝1 𝑞 𝑛−1 + ( ) 𝑝2 𝑞 𝑛−2 + ⋯ ⋯ ⋯ ⋯ ⋯ + ( ) 𝑝𝑛 𝑞 𝑛−𝑛
0 1 2 𝑛
= 𝑞 𝑛 + 𝑛𝐶1 𝑝1 𝑞 𝑛−1 + 𝑛𝐶2 𝑝2 𝑞 𝑛−2 + ⋯ ⋯ ⋯ ⋯ ⋯ + 𝑝𝑛
= (𝑞 + 𝑝)𝑛 = (1)𝑛 = 1
Thus the total density function of binomial distribution is unity. (proved).
#Problem 2: Find the moment generating function and cumulant generating function of binomial distribution.
Ans: The moment generating function of a random variable X about the origin is defined as
𝑀𝑥 (𝑡) = 𝐸(𝑒 𝑡𝑥 ) where 𝑡 is any arbitrary real value.
or, 𝑀𝑥 (𝑡) = ∑𝑛𝑥=0 𝑒 𝑡𝑥 𝑃(𝑥) [∵ 𝐸(𝑥) = ∑𝑛𝑥=0 𝑥 𝑃(𝑥)]
𝑛
= ∑𝑛𝑥=0 𝑒 𝑡𝑥 ( ) 𝑝 𝑥 𝑞 𝑛−𝑥
𝑥
𝑛 𝑛
= ∑𝑥=0(𝑝𝑒 )𝑥 ( ) 𝑞 𝑛−𝑥
𝑡
𝑥
𝑛 𝑛 𝑛 𝑛
= (𝑝𝑒 ) ( ) 𝑞 𝑛−0 + (𝑝𝑒 𝑡 )1 ( ) 𝑞 𝑛−1 + (𝑝𝑒 𝑡 )2 ( ) 𝑞 𝑛−2 + ⋯ ⋯ ⋯ + (𝑝𝑒 𝑡 )𝑛 ( ) 𝑞 𝑛−𝑛
𝑡 0
0 1 2 𝑛
= 𝑞 𝑛 + 𝑛𝐶1 𝑝𝑒 𝑡 𝑞 𝑛−1 + 𝑛𝐶2 (𝑝𝑒 𝑡 )2 𝑞 𝑛−2 + ⋯ ⋯ ⋯ + (𝑝𝑒 𝑡 )𝑛
∴ 𝑀𝑥 (𝑡) = (𝑞 + 𝑝𝑒 𝑡 )𝑛 which is the moment generating function of binomial distribution.
The cumulant generating function of binomial distribution is 𝐾𝑥 (𝑡) = 𝐿𝑜𝑔𝑀𝑥 (𝑡)
= 𝐿𝑜𝑔(𝑞 + 𝑝𝑒 𝑡 )𝑛
∴ 𝐾𝑥 (𝑡) = 𝑛𝐿𝑜𝑔(𝑞 + 𝑝𝑒 𝑡 )
Which is the cumulant generating function of binomial distribution.
Question 3: Find the mean and variance of binomial distribution.
Ans. We know that the moment generating function of binomial distribution is
𝑀𝑥 (𝑡) = (𝑞 + 𝑝𝑒 𝑡 )𝑛 .....................(i)
𝑑 𝑑
or, 𝑑𝑡 [𝑀𝑥 (𝑡)]𝑡=0 = 𝑑𝑡 [(𝑞 + 𝑝𝑒 𝑡 )𝑛 ]𝑡=0 [By differentiating with respect to t]
𝑡 𝑛−1 𝑡
= [𝑛(𝑞 + 𝑝𝑒 ) 𝑝𝑒 ]𝑡=0 ........................(ii)
∴ 𝑀𝑒𝑎𝑛 = 𝜇1′ = 𝑛(𝑞 + 𝑝𝑒 0 )𝑛−1 𝑝𝑒 0
= 𝑛(𝑞 + 𝑝)𝑛−1 𝑝
= 𝑛𝑝 [∵ 𝑝 + 𝑞 = 1]
Again differentiating (ii) with respect to t we get,
𝑑2 𝑑
[𝑀𝑥 (𝑡)]𝑡=0 = [𝑛(𝑞 + 𝑝𝑒 𝑡 )𝑛−1 𝑝𝑒 𝑡 ]𝑡=0 = [𝑛(𝑛 − 1)(𝑞 + 𝑝𝑒 𝑡 )𝑛−2 𝑝𝑒 𝑡 . 𝑝𝑒 𝑡 + 𝑛(𝑞 + 𝑝𝑒 𝑡 )𝑛−2 . 𝑝𝑒 𝑡 ]𝑡=0
𝑑𝑡 2 𝑑𝑡
𝑜𝑟, 𝜇2′ = [𝑛(𝑛 − 1)(𝑞 + 𝑝)𝑛−2 𝑝. 𝑝 + 𝑛(𝑞 + 𝑝)𝑛−2 . 𝑝]
𝑜𝑟, 𝜇2′ = [𝑛(𝑛 − 1)𝑝. 𝑝 + 𝑛. 𝑝] [∵ 𝑝 + 𝑞 = 1]
𝑜𝑟, 𝜇2′ = 𝑛2 𝑝2 − 𝑛𝑝2 + 𝑛𝑝
 Math 2225 (MHU) 2

We know, 𝜇2 = 𝜇2′ − (𝜇1′ )2

𝑜𝑟, 𝜇2 = 𝑛2 𝑝2 − 𝑛𝑝2 + 𝑛𝑝 − 𝑛2 𝑝2
= 𝑛𝑝(1 − 𝑝)
= 𝑛𝑝𝑞 This is required variance.
Problem 4: A symmetric die is thrown 600 times; find the binomial distribution getting 80 to 120.
Solution: Let X be the total number of successes from 𝑛 = 600 independent trials, probability of success in
1 5
each trial is 𝑃 = 𝑃(𝑠) = 6 and probability of failure in each trial is 𝑞 = 1 − 𝑝 = 6 thus X follows binomial
1
distribution. 𝜇 = 𝐸(𝑋) = 𝑛𝑝 = 600 × 6 = 100
1 5 500
and 𝜎 2 = 𝑉(𝑋) = 𝑛𝑝𝑞 = 600 × 6 × 6 = 6
500
∴𝜎=√ 6
1
By Chebyshave inequality, 𝑃{|𝑥 − 𝜇| ≤ 𝑘𝜎} ≥ 1 − 𝑘 2
1
or, 𝑃{−𝑘𝜎 < 𝑥 − 𝜇 < 𝑘𝜎} ≥ 1 − 𝑘 2
1
or, 𝑃{𝜇 − 𝑘𝜎 < 𝑥 < 𝜇 + 𝑘𝜎} ≥ 1 − 𝑘 2
500 500 1
or, 𝑃 {100 − 𝑘√ < 𝑥 < 100 + 𝑘√ } ≥ 1 − 𝑘2
6 6

1 500 20×√6
or, 𝑃{80 < 𝑥 < 120} ≥ 1 − 𝑘 2 [∵ 100 − 𝑘√ = 80 𝑜𝑟, 𝑘 = ]
6 √500
500
or, 𝑃{80 < 𝑥 < 120} ≥ 1 − 400×6
5
or, 𝑃{80 < 𝑥 < 120} ≥ 1 −
24
19
or, 𝑃{80 < 𝑥 < 120} ≥ 24Ans.
Problem 5: If 3% of the electric bulbs manufactured by a company are defective, find the probability
that in a sample of 100 bulbs (i) 3, (ii) at least 4, (iii) at most 2 will be defective.

Solution: Probability of defective bulbs= p = 3% =0.03

Probability of not defective bulbs = q = 1-p = 1- 0.03 = 0.97
Total number of bulbs = n = 100
(i)Probability of 3 defective bulbs = P(3) = 100𝐶3 (𝑝)3 (𝑞)100−3 = 100𝐶3 (0.03)3 (0.97)97
= 0.2275 Ans.
(ii) Probability of at least 4 defective = P(4)+P(5)+P(6)+ - --------- +P(100)
=1- [P(0)+P(1)+P(2)+P(3)]
Now, P(0) = 100𝐶0 (𝑝)0 (𝑞)100−0 = 100𝐶0 (0.03)0 (0.97)100 = 0.0476

P(1) = 100𝐶1 (𝑝)1 (𝑞)100−1 = 100𝐶1 (0.03)1 (0.97)99 = 0.1471

P(2) = 100𝐶2 (𝑝)2 (𝑞)100−2 = 100𝐶2 (0.03)2 (0.97)98 = 0.2252

So, Probability of at least 4 defective=1- [0.0476+0.1471+0.2252+0.2275]= 0.3526 Ans.

(iii) Probability of at most 2 defective = P (0 or 1 or 2) = P(0)+P(1)+P(2)
= 0.0476+0.1471+0.2252=0.4199 Ans.
H.W. Problem 6: Seventy percent signals sent from a server reach to its goal properly. Once 3 signals
are checked randomly. Find the probability that out of 3, (a) at least 1 reaches properly, and (b) at best
2 reaches properly.
 Math 2225 (MHU) 3

Normal Distribution
#Question 1: Define normal distribution. Prove that, the total probability of normal distribution is unity.
Normal Distribution: A continuous random variable 𝑋 is said to be a normal distribution with parameter𝜇
and 𝜎 2 if its density function is given by
1 𝑥−𝜇 2
1 − ( )
𝑓(𝑥) = 𝜎√2𝜋 𝑒 2 𝜎 ; −∞ < 𝑥 < ∞, −∞ < 𝜇 < ∞,

Which is expressed by 𝑋~N(𝜇, 𝜎 2 )

1 𝑥−𝜇 2
1
Proof: We know, 𝑓(𝑥) = 𝜎√2𝜋 𝑒 −2( )
𝜎 ; −∞ < 𝑥 < ∞, −∞ < 𝜇 < ∞
1 𝑥−𝜇 2
∞ ∞ 1 − ( )
∴ 𝑃(−∞ < 𝑥 < ∞) = ∫−∞ 𝑓(𝑥) 𝑑𝑥 = ∫−∞ 𝜎√2𝜋 𝑒 2 𝜎 𝑑𝑥
𝑉2
1 ∞ 𝑥−𝜇
= 𝜎√2𝜋 ∫−∞ 𝑒 − 2 𝜎 𝑑𝑉 𝐿𝑒𝑡, 𝑉 = ,
𝜎
𝑉2
1 ∞ −
= ∫ 𝑒 2 𝑑𝑉 ∴ 𝑥 = 𝜇 + 𝜎𝑉
√2𝜋 −∞

𝑜𝑟, 𝑑𝑥 = 𝜎𝑑𝑉, 𝑤ℎ𝑒𝑛, −∞ < 𝑥 < ∞

𝑡ℎ𝑒𝑛 − ∞ < 𝑉 < ∞
𝑉2
1 ∞ − 𝑉2 𝑑𝑧
= ∫ 𝑒 2 𝑑𝑉 𝐿𝑒𝑡, =𝑧 𝑜𝑟, 𝑉 = √2𝑧, 𝑜𝑟, 𝑑𝑉 = ,
√2𝜋 −∞ 2 √2𝑧
∞ 1
2 − 𝑑𝑧
= ∫ 𝑒 −𝑧 𝑧 2 √2
√2𝜋 0
∞ 1
1 −1
= ∫ 𝑒 −𝑧 𝑧 2 𝑑𝑧
√𝜋 0

1 1
=
 2
1
= √𝜋 =1
√𝜋

Thus, the total probability of normal distribution is unity. (proved).

Question 2: Find the mean and variance of normal distribution.
1 𝑥−𝜇 2
∞ ∞ 1
Ans. We know, 𝐸(𝑥) = ∫−∞ 𝑥𝑓(𝑥) 𝑑𝑥 = ∫−∞ 𝑥 𝜎√2𝜋 𝑒 −2( )
𝜎 𝑑𝑥
𝑉2
1 ∞ − 𝑥−𝜇
= ∫ (𝜇 + 𝜎𝑉)𝑒 2 𝜎 𝑑𝑉 (𝐿𝑒𝑡, 𝑉 = ,
𝜎√2𝜋 −∞ 𝜎

𝑉2 𝑉2 ∴ 𝑥 = 𝜇 + 𝜎𝑉
1 ∞ − 1 ∞ −
= ∫ 𝜇𝑒
√2𝜋 −∞
2 𝑑𝑉 + ∫ 𝜎𝑉𝑒
√2𝜋 −∞
2 𝑑𝑉 (𝑜𝑟, 𝑑𝑥 = 𝜎𝑑𝑉, 𝑤ℎ𝑒𝑛, −∞ < 𝑥 < ∞
𝑡ℎ𝑒𝑛 − ∞ < 𝑉 < ∞
2 2
𝜇 ∞ −𝑉 𝜎 ∞ −
𝑉
= ∫ 𝑒 2 𝑑𝑉 + ∫ 𝑉𝑒 2 𝑑𝑉
√2𝜋 −∞ √2𝜋 −∞
 Math 2225 (MHU) 4
𝑉2
𝐿𝑒𝑡, = 𝑧, 𝑜𝑟, 𝑉 = √2𝑧,
2
𝑑𝑧
𝜇 ∞ −𝑧 𝑑𝑧 𝜎 ∞ −𝑧 𝑑𝑧 𝑜𝑟, 𝑑𝑉 = ,
= ∫−∞
𝑒 + ∫−∞
√2𝑧𝑒 √2𝑧
√2𝜋 √2𝑧 √2𝜋 √2𝑧
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
∞ 𝑉2
−
(∫−∞ 𝑉𝑒 2 𝑑𝑉 = 0 𝑓𝑜𝑟 𝑜𝑑𝑑 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
1
𝜇 ∞ − 2𝜎 ∞
= ∫ 𝑒 −𝑧 𝑧 2 𝑑𝑧 + √2𝜋 ∫0 𝑒 −𝑧 𝑧 0 𝑑𝑧
√𝜋 0

 1 2
= + .0
 2 2
𝜇
= √𝜋
√𝜋

∴ 𝐸(𝑋) = 𝜇 =Mean.
2nd part: We know, 𝑉(𝑋) = 𝐸(𝑋 − 𝜇)2 ...............(i)
1 𝑥−𝜇 2
∞ ∞ 1
Now, 𝐸(𝑋 − 𝜇)2 = ∫−∞(𝑥 − 𝜇)2 𝑓(𝑥) 𝑑𝑥 = ∫−∞(𝑥 − 𝜇)2 𝜎√2𝜋 𝑒 −2( )
𝜎 𝑑𝑥
𝑣 2
𝜎2∞ 2 −2 𝑥−𝜇
= ∫ 𝑣 𝑒 𝑑𝑥 (𝐿𝑒𝑡, 𝑣 = ⇒ 𝑥 − 𝜇 = 𝑣𝜎 ∴ 𝑑𝑥 = 𝜎𝑑𝑣,
√2𝜋 −∞ 𝜎
𝜎2 ∞ 𝑑𝑧 𝑣2 𝑑𝑧
= ∫ 2𝑧𝑒 −𝑧
√2𝜋 −∞ √2𝑧
(𝐿𝑒𝑡, 2
= 𝑧, 𝑜𝑟, 𝑣 = √2𝑧, 𝑜𝑟, 𝑑𝑣 =
√2𝑧
1
2𝜎2 ∞
= ∫0 𝑒 −𝑧 𝑧 𝑑𝑧
2
√𝜋
3
2𝜎2 ∞ −1
= ∫0 𝑒 −𝑧 𝑧 2 𝑑𝑧
√𝜋

2 2 3
=
 2

2 2  3  3
=  − 1 − 1
 2  2

2 2 1 1
=
 2 2
2𝜎2 1
= √𝜋
√𝜋 2

= 𝜎 2 = Variance.
2nd part: Or Rule:
We know, 𝑉(𝑋) = 𝐸(𝑋 2 ) − {𝐸(𝑋)}2 ...............(i)
1 𝑥−𝜇 2
∞ ∞ 1
Now, 𝐸(𝑋 2 ) = ∫−∞ 𝑥 2 𝑓(𝑥) 𝑑𝑥 = ∫−∞ 𝑥 2 𝜎√2𝜋 𝑒 −2( )
𝜎 𝑑𝑥
𝑣2
1 ∞ 𝑥−𝜇
= 𝜎√2𝜋 ∫−∞(𝜇 + 𝜎𝑣)2 𝑒 − 2 𝜎 𝑑𝑣 (𝐿𝑒𝑡, 𝑣 = , ∴ 𝑥 = 𝜇 + 𝜎𝑣
𝜎

 v2
1 − 𝑜𝑟, 𝑑𝑥 = 𝜎𝑑𝑣, 𝑤ℎ𝑒𝑛, −∞ < 𝑥 < ∞
=  ( + 2v +  v )e  dv
2 2 2 2
(
 2 −
𝑡ℎ𝑒𝑛 − ∞ < 𝑣 < ∞
𝑣2 𝑣2 𝑣2
𝜇2 ∞ − 2𝜇𝜎 ∞ − 𝜎2 ∞ −
= ∫ 𝑒
√2𝜋 −∞
2 𝑑𝑣 + ∫ 𝑣𝑒
√2𝜋 −∞
2 𝑑𝑣 + ∫ 𝑣 2𝑒
√2𝜋 −∞
2 𝑑𝑣
 Math 2225 (MHU) 5
  
2 dz 2 dz 2 dz 𝑣2
e   2 ze
−z
= + 2ze−z + −z
(𝐿𝑒𝑡, = 𝑧, 𝑜𝑟, 𝑣 = √2𝑧,
2 − 2z 2 − 2z 2 − 2z 2

1 1
𝜇2 ∞ − 4𝜇𝜎 ∞ 2𝜎2 ∞ 𝑑𝑧
= ∫ 𝑒 −𝑧 𝑧 2 𝑑𝑧 + √2𝜋 ∫0 𝑒 −𝑧 𝑧 0 𝑑𝑧 + ∫0 𝑒 −𝑧 𝑧 2 𝑑𝑧 (𝑜𝑟, 𝑑𝑣 =
√𝜋 0 √𝜋 √2𝑧

 2 1 2 2 2 − z 2 −1
3

 0
= + .0 + e z dz
 2 2

2 2 2 3
= +
  2

2 2 2  3  3
= +  − 1 − 1
  2  2

2 2 2 1 1
= +
  2 2
𝜇2 2𝜎2 1
= √𝜋 + √𝜋
√𝜋 √𝜋 2

= 𝜇2 + 𝜎 2
Hence from (i) we get, 𝑉(𝑋) = 𝐸(𝑋 2 ) − {𝐸(𝑋)}2
= 𝜇2 + 𝜎 2 − 𝜇2
= 𝜎 2 = Variance.

Question 3: Find the moment generating function, Skewness and kurtosis and comment on the shape.
1 𝑥−𝜇 2
∞ ∞ 1
Ans.: We know, 𝑀𝑋 (𝑡) = 𝐸(𝑒 𝑡𝑥 ) = ∫−∞ 𝑒 𝑡𝑥 𝑓(𝑥) 𝑑𝑥 = ∫−∞ 𝑒 𝑡𝑥 𝜎√2𝜋 𝑒 −2( )
𝜎 𝑑𝑥
2
1 ∞ 𝑡(𝜇+𝜎𝑣) −𝑣 𝑥−𝜇
= ∫ 𝑒 𝑒 2 𝜎 𝑑𝑣 (𝐿𝑒𝑡, 𝑣 = ,
𝜎√2𝜋 −∞ 𝜎

𝑣2 ∴ 𝑥 = 𝜇 + 𝜎𝑣
1 ∞ −
= ∫ 𝑒 𝑡𝜇 𝑒 𝑡𝑣𝜎 𝑒
√2𝜋 −∞
2 𝑑𝑣 (𝑜𝑟, 𝑑𝑥 = 𝜎𝑑𝑣, 𝑤ℎ𝑒𝑛, −∞ < 𝑥 < ∞
𝑡ℎ𝑒𝑛 − ∞ < 𝑣 < ∞
1
𝑒 𝑡𝜇 ∞ − (𝑣 2 −2𝑡𝑣𝜎)
= ∫ 𝑒 2 𝑑𝑣
√2𝜋 −∞
1
𝑒 𝑡𝜇 ∞ − (𝑣 2 −2𝑣.𝑡𝜎+𝑡 2 𝜎2 −𝑡 2 𝜎2 )
= ∫ 𝑒 2 𝑑𝑣
√2𝜋 −∞
𝑡2 𝜎2
𝑡𝜇+ 1
𝑒 2 ∞ − (𝑣−𝑡𝜎)2
= ∫−∞ 𝑒 2 𝑑𝑣 (𝐿𝑒𝑡, 𝑣 − 𝑡𝜎 = 𝑧, 𝑜𝑟, 𝑑𝑣 = 𝑑𝑧,
√2𝜋
𝑡2 𝜎2 2
𝑡𝜇+
𝑒 2 ∞ −𝑧 𝑧2 𝑑𝑦
= ∫−∞ 𝑒 2 𝑑𝑧 (𝐿𝑒𝑡, = 𝑦, 𝑜𝑟, 𝑧 = √2𝑦, 𝑜𝑟, 𝑑𝑧 = )
√2𝜋 2 √2𝑦

𝑡2 𝜎2
𝑡𝜇+
𝑒 2 ∞ 𝑑𝑦
=
√2𝜋
∫−∞ 𝑒 −𝑦 √2𝑦
𝑡2 𝜎2
𝑡𝜇+ 1
𝑒 2 ∞
= 2 ∫0 𝑒 −𝑦 𝑦 −2 𝑑𝑦
2√𝜋
 Math 2225 (MHU) 6
𝑡2 𝜎2
𝑡𝜇+ 1
𝑒 2 ∞ −1
= ∫0 𝑒 −𝑦 𝑦 2 𝑑𝑦
√𝜋

t 2 2
t +
2
e 1
=
 2
𝑡2 𝜎2
𝑡𝜇+
𝑒 2
= √𝜋
√𝜋
𝑡2 𝜎2
= 𝑒 𝜇𝑡+ 2 , which is the moment generating function of normal distribution.
Cumulant generating function:
𝑡2 𝜎2
𝑡 2 𝜎2
𝑘𝑥 (𝑡) = 𝐿𝑜𝑔𝑀𝑥 (𝑡) = 𝐿𝑜𝑔𝑒 𝜇𝑡+ 2 = 𝜇𝑡 + 2
𝑡 𝑡2 𝑡3 𝑡4
= 𝜇 + 𝜎 2 + 0. + 0. + ⋯ ⋯ ⋯
1! 2! 3! 4!
𝑡
∴ 𝑘1 =Co-efficient of 1! in 𝑘𝑥 (𝑡) = 𝜇 = 𝑀𝑒𝑎𝑛
𝑡2
𝑘2 =Co-efficient of in 𝑘𝑥 (𝑡) = 𝜎 2 = 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒
2!
𝑡3
𝑘3 =Co-efficient of in 𝑘𝑥 (𝑡) = 0
3!
𝑡4
𝑘4 =Co-efficient of in 𝑘𝑥 (𝑡) = 0
4!

Now, 𝜇1′ = 𝑘1 = 𝜇
𝜇2 = 𝑘2 = 𝜎 2
𝜇3 = 𝑘3 = 0
𝜇4 = 𝑘4 + 3𝑘22 = 0 + 3𝜎 4 = 3𝜎 4
𝜇2 0
So, the Skewness, 𝛽1 = 𝜇33 = (𝜎2 )3 = 0.
2

Comment: Since Skewness,𝛽1 = 0, so, the normal distribution is symmetric.

𝜇 3𝜎4
Also the kurtosis, 𝛽2 = 𝜇42 = =3
2 𝜎4

Comment: Since kurtosis, 𝛽2 = 3, So, the normal distribution is mesokurtic.

Question 4: Proved that mean, median and mode of normal distribution is coincide.
1 𝑥−𝜇 2
1
Proof: We know, 𝑓(𝑥) = 𝜎√2𝜋 𝑒 −2( )
𝜎 ; −∞ < 𝑥 < ∞

Here, mean = 𝜇 and variance= 𝜎 2

𝑀
Let, 𝑀be the median, By definition we can write, ∫−∞ 𝑓(𝑥)𝑑𝑥 = 0.5
𝜇 𝑀
𝑜𝑟, ∫−∞ 𝑓(𝑥)𝑑𝑥 + ∫𝜇 𝑓(𝑥)𝑑𝑥 = 0.5
𝑀
𝑜𝑟, 0.5 + ∫𝜇 𝑓(𝑥)𝑑𝑥 = 0.5
𝑀
𝑜𝑟 ∫𝜇 𝑓(𝑥)𝑑𝑥 = 0,

∴𝑀=𝜇
Hence, median=mean.....................(i)
 Math 2225 (MHU) 7
1 𝑀0−𝜇 2
1
Again, let, 𝑀0 be the mode, we get, 𝑓(𝑀0 ) = 𝜎√2𝜋 𝑒 −2( )
𝜎 ; −∞ < 𝑀0 < ∞..............(ii)

Now, Mode is the solution of 𝑓 ′ (𝑀0 ) = 0 & 𝑓′′((𝑀0 ) < 0

1 𝑀0−𝜇 2
1
From (ii), 𝐿𝑜𝑔 𝑓(𝑀0 ) = 𝐿𝑜𝑔 𝜎√2𝜋 𝑒 −2( )
𝜎

1 𝑀 −𝜇 2
1 − ( 0 )
or, 𝐿𝑜𝑔 𝑓(𝑀0 ) = 𝐿𝑜𝑔 𝜎√2𝜋 + 𝐿𝑜𝑔𝑒 2 𝜎

1 1 𝑀0 −𝜇 2
or, 𝐿𝑜𝑔 𝑓(𝑀0 ) = 𝐿𝑜𝑔 𝜎√2𝜋 − 2 ( )
𝜎
1 𝑀0 −𝜇 1
or, 𝑓 ′ (𝑀0 ) = 0 − ( ) .𝜎 [By differentiating with respect to 𝑀0 ]
𝑓(𝑀0 ) 𝜎
𝑀0 −𝜇
or, 𝑓 ′ (𝑀0 ) = −𝑓(𝑀0 ) ( )
𝜎2
𝑀0 −𝜇 𝑀 −𝜇 1
or, 𝑓(𝑀0 ) ( ) = 0, [∵ 𝑓′(𝑀0 ) = 0] Now, 𝑓 ′′ (𝑀0 ) = 𝑓 ′ (𝑀0 ) ( 𝜎0 2 ) − 𝑓(𝑀0 ) 𝜎2
𝜎2
At 𝑀0 = 𝜇,
or, 𝑀0 − 𝜇 = 0, 1 1
𝑓 ′′ (𝑀0 ) = −𝑓(𝑀0 ) 𝜎2 and 𝑓(𝑀0 ) = 𝜎√2𝜋 ;
∴ 𝑀0 = 𝜇.
Hence, Mean=Mode.......(iii) 1 1
∴ 𝑓 ′′ (𝑀0 ) = − <0
𝜎√2𝜋 𝜎 2
Therefore from (i) and (iii) we get, mean=median=mode,
i.e; mean, median and mode of normal distribution is coincide. Proved.
Question 5: What is the chief characteristic of normal distribution/ property of normal distribution.
Answer: The following property holds normal distribution:
(i) The curve is bell shaped and symmetrical about the line 𝑥 = 𝜇.
(ii) mean, median and mode of the distribution coincide.
(iii) Odd central moment, 𝜇2𝑟+1 = 0 and even central moment, 𝜇2𝑟 = 1,3,5, ⋯ ⋯ ⋯ (2𝑟 − 1)𝜎 2𝑟 .
(iv) 𝛽1 = 0 and 𝛽2 = 3.
(v) 𝑥 - axis is an symmetric to the curve.

Question 6: X is normally distributed and the mean of X is 12 and standard deviation is 4, find the following
probability: (i) 𝑋 ≥ 20 (ii) 𝑋 ≤ 20 and (iii) 0 ≤ 𝑋 ≤ 12.
Solution: We have, 𝜇 = 12, 𝜎 = 4,
i.e; 𝑋~N(𝜇, 𝜎 2 )
or, 𝑋~N(12, 16)
𝑋−𝜇
(i) 𝑃(𝑋 ≥ 20) =? Let, 𝑧 = ,
𝜎
∴ 𝑃(𝑋 ≥ 20) = 𝑃(𝑧 ≥ 2) 20−12
or, 𝑧 = when 𝑋 = 20, 𝜇 = 12, 𝜎 = 4.
4
= 𝑃(0 < 𝑧 < ∞) − 𝑃(0 ≤ 𝑧 ≤ 2)
or, 𝑧 = 2
=0.5-0.4772
=0.0228 Ans.
(ii) 𝑃(𝑋 ≤ 20) = 1 − 𝑃(𝑋 ≥ 20)
0 2
=1- 0.0228
 Math 2225 (MHU) 8

=0.9772 Ans. 𝑋−𝜇

Let, 𝑧 = ,
𝜎
0−12
(iii) 𝑃(0 ≤ 𝑋 ≤ 12) = 𝑃(−3 ≤ 𝑧 ≤ 0) or, 𝑧 = = −3 when 𝑋 = 0, 𝜇 = 12, 𝜎 = 4.
4
12−12
=𝑃(0 ≤ 𝑧 ≤ 3) and 𝑧 = = 0 when 𝑋 = 12, 𝜇 = 12, 𝜎 = 4.
4
=0.4987 Ans.

Question 7: A random variable follows the normal probability distribution with mean 80 and a
standard deviation of 20. What is the probability that a randomly selected value from this population
(i) is less than 110 and (ii) is more than 40.

Solution: We have, 𝜇 = 80, 𝜎 = 20,

i.e; 𝑋~N(𝜇, 𝜎 2 )
or, 𝑋~N(80, 400)
𝑋−𝜇
(i) 𝑃(𝑋 ≤ 110) = 1 − 𝑃(𝑋 ≥ 110) Let, 𝑧 = ,
𝜎
Now 𝑃(𝑋 ≥ 110) = 𝑃(𝑧 ≥ 1.5) 110−80
or, 𝑧 = when 𝑋 = 110, 𝜇 = 80, 𝜎 = 20.
20
= 𝑃(0 < 𝑧 < ∞) − 𝑃(0 ≤ 𝑧 ≤ 1.5)
or, 𝑧 = 1.5
=0.5-0.4332
=0.0668
∴ 𝑃(𝑋 ≤ 110) = 1 − 𝑃(𝑋 ≥ 110)
0 1.5
= 1 − 0.0668
= 0.9332 Ans.
𝑋−𝜇
(ii) 𝑃(𝑋 ≥ 40) = 𝑃(𝑧 ≥ −2) Let, 𝑧 = ,
𝜎
40−80
=1- 𝑃(𝑧 ≤ −2) or, 𝑧 = = −2 when 𝑋 = 40, 𝜇 = 80, 𝜎 = 20
20
= 1 − {𝑃(−∞ < 𝑧 < 0) − 𝑃(−2 ≤ 𝑧 ≤ 0)}
= 1 − {𝑃(0 < 𝑧 < ∞) − 𝑃(0 ≤ 𝑧 ≤ 2)}
= 1 − {0.5 − 0.4772}
= 1 − 0.0228
= 0.9772 Ans.

[ N.B.: Statistics chart: Area under standard normal curve from 0 to Z.]
 Math 2225 (MHU) 9

Poisson Distribution
#Question 1: Define Poisson distribution. Show that Poisson distribution is a particular limiting form of the
Binomial distribution.
𝑚𝑟 𝑒 −𝑚
Definition: Poisson distribution is 𝑃(𝑟) = where 𝑚is the mean of the distribution.
𝑟!
Proof: Poisson distribution is a particular limiting form of the Binomial distribution when 𝑝 (or, 𝑞) is very
small and 𝑛 is large enough.
In Binomial distribution, 𝑃(𝑟) = 𝑛𝐶𝑟 𝑞𝑛−𝑟 𝑝𝑟
= 𝑛𝐶𝑟 (1 − 𝑝)𝑛−𝑟 𝑝𝑟 [∵ 𝑝 + 𝑞 = 1]
𝑚 𝑛−𝑟 𝑚 𝑟 𝑚
= 𝑛𝐶𝑟 (1 − ) ( ) (Since 𝑚 = 𝑛𝑝 ∴𝑝= )
𝑛 𝑛 𝑛

𝑛! 𝑚 𝑛−𝑟 𝑚 𝑟
= (1 − ) ( )
𝑟!(𝑛−𝑟)! 𝑛 𝑛

𝑛(𝑛−1)(𝑛−2)⋯⋯(𝑛−𝑟−1) 𝑚 𝑟 𝑚 𝑛−𝑟
= ( ) (1 − )
𝑟! 𝑛 𝑛
𝑛 𝑛−1 𝑛−2 𝑛−𝑟−1 𝑚 𝑛
( )( )⋯⋯( )(𝑚)𝑟 (1− )
𝑛 𝑛 𝑛 𝑛 𝑛
= 𝑚 𝑟
𝑟!(1− )
𝑛

1 2 𝑟−1 𝑚 𝑛
(1− )(1− )⋯⋯(1− )𝑚𝑟 (1− )
𝑛 𝑛 𝑛 𝑛
= 𝑚 𝑟
𝑟!(1− )
𝑛

Now, taking limit when n tends to infinity

−𝑛 −𝑚
𝑚 𝑛 𝑚 𝑚
lim (1 − 𝑛 ) = lim [(1 − ) ] = 𝑒 −𝑚
𝑛→∞ 𝑛→∞ 𝑛
𝑚𝑟 𝑒 −𝑚
∴ 𝑃(𝑟) = which is Poisson distribution
𝑟!
Hence, Poisson distribution is a particular limiting form of the Binomial distribution. (Proved).

Question 2: Find the mean and standard deviation of Poisson distribution.

Ans. In Binomial distribution, Mean: 𝑚 = 𝑛𝑝
𝑚 𝑚
So, In Poisson distribution, Mean: 𝑚 = lim 𝑛𝑝 = lim 𝑛. 𝑛 = lim 𝑚 = 𝑚 (Since 𝑚 = 𝑛𝑝 ∴ 𝑝 = 𝑛)
𝑛→∞ 𝑛→∞ 𝑛→∞
2
2nd Part: In Binomial distribution, Variance: 𝜎 = 𝑛𝑝𝑞
So, In Poisson distribution, Variance: 𝜎 2 = lim 𝑛𝑝𝑞 = lim 𝑛𝑝(1 − 𝑝)
𝑛→∞ 𝑛→∞
𝑚 𝑚 𝑚
= lim 𝑛. 𝑛 (1 − 𝑛 ) = lim 𝑚 (1 − 𝑛 ) = 𝑚.
𝑛→∞ 𝑛→∞

Hence, the standard deviation of Poisson distribution is 𝜎 = √𝑚.

Problem 1: If the variance of the Poisson distribution is 2, Find the probabilities for 𝑟 = 1, 2, 3,4 from the
recurrence relation of the Poisson distribution. Also find 𝑃(𝑟 ≥ 4).
Solution: Given that Variance of Poisson distribution: 𝜎 2 = 𝑚 = 2
So, mean of Poisson distribution: = 𝑚 = 2.
𝑚𝑟 𝑒 −𝑚 2𝑟 𝑒 −2
We know, Poisson distribution is 𝑃(𝑟) = = .
𝑟! 𝑟!
20 𝑒 −2
∴ 𝑃(0) = = 𝑒 −2 = 0.1355
0!
 Math 2225 (MHU) 10
21 𝑒 −2
𝑃(1) = = 0.271
1!
22 𝑒 −2
𝑃(2) = = 0.271
2!
23 𝑒 −2
𝑃(3) = = 0.180
3!
24 𝑒 −2
𝑃(4) = = 0.090
4!

∴ 𝑃(𝑟 ≥ 4) = 𝑃(4) + 𝑃(5) + 𝑃(6) + 𝑃(7) + ⋯ ⋯

= 𝑃(0) + 𝑃(1) + 𝑃(2) + 𝑃(3) + 𝑃(4) + 𝑃(5) + 𝑃(6) + 𝑃(7) + ⋯ ⋯ − [𝑃(0) +
𝑃(1) + 𝑃(2) + 𝑃(3)]
= 1 − [𝑃(0) + 𝑃(1) + 𝑃(2) + 𝑃(3)]
= 1 − [0.135 + 0.271 + 0.271 + 0.180]
= 1 − 0.857
= 0.143 Ans.

Problem 2: The probability of breaking a glass breaker while heating in a laboratory is 0.012. If we heat 200
such glass breakers, find the probability of breaking exactly 2 beakers using: i) Binomial distribution ii)
Poisson distribution.
Solution: Given that 𝑛 = 200,
Success event, 𝑝 = 0.012,
Required probability, x= 2
i) In Binomial distribution, we know, 𝑝 + 𝑞 = 1
Or, 𝑞 = 1 − 𝑝 = 1 − 0.012 = 0.988,
𝑛
We know, the Binomial distribution is 𝑃(𝑥) = ( ) 𝑝 𝑥 𝑞 𝑛−𝑥
𝑥
200
Or, 𝑃(2) = ( ) (0.012)2 (0.988)(200−2)
2
= 19900 × (0.012)2 (0.988)198
= 0.262477 Ans.
ii) In Poisson distribution, we know, 𝑚 = 𝑛𝑝 = 200 × 0.012 = 2.4
𝑚𝑥 𝑒 −𝑚
We know, the Poisson distribution is 𝑃(𝑥) =
𝑥!
2
(2.4) 𝑒−2.4
Or, 𝑃(2) = 2!
= 0.261268 Ans.

Problem 3: Suppose 3% of bolts made by a machine are defective, the defects occurring at random during
production. If bolts are packaged 50 per box, find (a) exact probability and (b) Poisson approximation to it,
that a given box will contain 5 defectives. Ans. (a) 0.013074, (b) 0.01412

Problem-4: Using Poisson distribution, find the probability that the ace of spades will be drawn from a pack
of well-shuffled cards at least once in 104 consecutive trails. Ans. 0.865