You are on page 1of 14

Prof Iyer APSL(IIM-L) Puhlished : 4/29/2021

ib.mathfaculty@gmail.com

This output is developed keeping in mind

- Support to Unit 4 Paper 3 Mathematical Investigation for


HL students of AA and AI
- An attempt is made to analyze the probability density function of Normal distribution
- An output to help the Math exploration for students of HL of both AA and AI

Next in the series :


Applying Calculus in Economics
Curriculum Extension : Partial derivatives , Unconstrained and Constrained optimizations
and Lagrange Multipliers

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Applying Calculus in Normal Distribution

Developed by From Prof Iyer


Part of my Paper 3 Inputs for HL-AA and AI /IA inputs for HL
______________________________________________________________________________

Theory:

1. Mean = E(X) = ∫−∞ 𝒙 𝒇(𝒙)𝒅𝒙
∞ ∞ 𝟐
2. Var (x)= ∫−∞[𝒙𝟐 𝒇(𝒙)𝒅𝒙 − (∫−∞ 𝒙 𝒇(𝒙)𝒅𝒙 ) ]
𝑴 𝟏
3. Median = ∫−∞ 𝒇(𝒙)𝒅𝒙 =
𝟐
′ (𝒎)
4. Mode = 𝒇 = 𝟎 𝒂𝒏𝒅 𝒇′′ (𝒎) < 𝟎 , 𝒕𝒉𝒆𝒏 𝒎 𝒊𝒔 𝒕𝒉𝒆 𝒎𝒐𝒅𝒆
4. Probability density function (pdf) of Normal distribution
(𝒙−𝝁)𝟐
𝟏 −
f(x) = 𝒆 𝟐𝝈𝟐 where 𝝁 𝒂𝒏𝒅 𝝈 are mean and sd respectively
𝝈√𝟐𝝅

______________________________________________________________________________

This work is not copied from any past paper nor from any text book. This is prepared to
understand the linkage of Calculus in Non-Calculus sections of the curriculum. .© 𝑷𝒓𝒐𝒇. 𝑰𝒚𝒆𝒓
Any criticism, comments or corrections, most welcome
Write to ib.mathfaculty@gmail.com

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Finding Expectation using the method of Calculus
(𝒙−𝝁)𝟐
∞ 𝟏 −
E(X) = ∫−∞ 𝒙 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐 𝒅𝒙

(𝒙−𝝁)𝟐
𝟏∞ −
 E(X) = ∫ 𝑥𝒆 𝟐𝝈𝟐 𝑑𝑥
𝝈√𝟐𝝅 −∞

 Using Integration by substitution

𝑥−𝜇
 Let = 𝑡 𝑎𝑛𝑑 𝜎𝑑𝑡 = 𝑑𝑥 substituting in the above integral
𝜎
Also when x =−∞ 𝑎𝑛𝑑 𝑡 = −∞ 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥 = ∞ , 𝑡 = ∞

𝒕𝟐
𝟏∞ −
 E(X) = ∫ (𝜎𝑡 + 𝜇)𝒆 𝟐 𝜎𝑑𝑡
𝝈√𝟐𝝅 −∞

𝒕𝟐 𝒕𝟐
𝝈 ∞ 𝝁𝜎 ∞
 E(X) = 𝝈√𝟐𝝅 ∫−∞(𝑡)𝒆− 𝟐 𝑑𝑡 + 𝝈√𝟐𝝅 ∫−∞ 𝒆− 𝟐 𝑑𝑡

𝒕 𝟐 𝟐
𝝈 ∞ − ∞ −𝒕
𝝁
 E(X) = ∫ (𝑡)𝒆 𝟐 𝑑𝑡 + ∫ 𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞ √𝟐𝝅 −∞

𝜎 𝝁
 E(X)= I1 + I2 (Let us assume ) ….(1)
√𝟐𝝅 √𝟐𝝅

∞ 𝒕𝟐
𝜎 𝜎 −
Let us Solve
√𝟐𝝅
I1 = ∫ (𝒕)𝒆 𝟐 𝒅𝒕
√𝟐𝝅 −∞

𝒕𝟐 t2
2t
 Put 𝒆− 𝟐 =u and du = e− 2 (− 2 ) dt = du

 Change of limits: - Also, when t =−∞ 𝑎𝑛𝑑 𝑢 = 0 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑡 = ∞ , 𝑢 = 0

t2
 So e− 2 (t)dt= - du

𝒕𝟐
𝜎 ∞ 𝝈 0
 ∫−∞
(𝑡)𝒆− 𝟐 𝑑𝑡 = ∫ −𝑑𝑢 = 0 ……..(2)
√𝟐𝝅 √𝟐𝝅 0

𝟐
𝝁 ∞ −𝒕 𝝁
Let us Solve I 2 = ∫ 𝒆 𝟐 𝒅𝒕
𝝈√𝟐𝝅 √𝟐𝝅 −∞

(𝒕−𝟎)𝟐 (𝒙−𝟎)𝟐
𝝁 ∞ − 𝟏 ∞ −
∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑡 = 𝜇{ ∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑥 }
√𝟐𝝅 √𝟐𝝅

Using property of integral variables can be changed with same limits.

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


This is a standard normal distribution
X ~𝑵 (𝝁, 𝝈𝟐 ) 𝑏𝑒𝑐𝑜𝑚𝑛𝑔 𝑿~𝑵(𝟎, 𝟏) with

(𝒙−𝟎)𝟐
𝟏 ∞ −
=𝝁( ∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑥 ) =𝜇(1)
√𝟐𝝅

Area under the curve for a standard normal distribution =1 With 𝜇 = 0 𝑎𝑛𝑑 𝜎 = 1

Alternate Method:
Using the iterative or arbitrating values of lower and upper limit or higher values
ie I2→ 1 as a→ ∞ for large values -a on LHS and +a on RHS of x -axis

(𝒙−𝟎)𝟐
𝟏 ∞ −
∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑥 ) = 1……… (3)
𝝈√𝟐𝝅

Therefore E(x) = 0 + 𝜇(1) = 𝜇 Substituting (3) and (2) in (1)

Proved

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Finding Variance using the method of Calculus
(𝒙−𝝁)𝟐 2
∞ 𝟏 − ∞
Var (X) = [∫−∞ 𝒙𝟐 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐 𝒅𝒙 ] − (∫−∞ 𝑥 𝑓(𝑥)𝑑𝑥 )

(𝒙−𝝁)𝟐
𝟏 ∞ −
 Var (X) = ∫ 𝒙𝟐 𝒆 𝟐𝝈𝟐
𝝈√𝟐𝝅 −∞
𝒅𝒙 - E(X)2
(𝒙−𝝁)𝟐
𝟏 ∞ 𝟐 −
 Var (X) = ∫−∞
𝒙 𝒆 𝟐𝝈𝟐 𝒅𝒙 - 𝝁2
𝝈√𝟐𝝅

 Using Integration by substitution for the integral section of the Var (x)

(𝒙−𝝁)𝟐
𝟏 ∞ −
 = 𝝈√𝟐𝝅 ∫−∞ 𝒙𝟐 𝒆 𝟐𝝈𝟐 𝒅𝒙

𝑥−𝜇
 Let = 𝑡 𝑎𝑛𝑑 𝜎𝑑𝑡 = 𝑑𝑥 and dx =𝜎𝑑𝑡
𝜎

 substituting in the above integral


Change of Limits: - When x =−∞ 𝑎𝑛𝑑 𝑡 = −∞ 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥 = ∞ , 𝑡 = ∞

∞ 𝒕𝟐
𝟏 −
 ∫ (𝜎𝑡 + 𝜇)2 𝒆 𝟐 (𝜎𝑑𝑡)
𝝈√𝟐𝝅 −∞

𝒕 𝟐 𝟐 𝟐
𝝈𝟐 ∞ 2 )𝒆− 𝟐 𝝁𝟐 𝝈 ∞ −𝒕 2𝜇𝜎 ∞ −
𝒕
 ∫ (𝑡 𝑑𝑡 + ∫ 𝒆 𝟐 𝑑𝑡 + ∫ 𝒕𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞ 𝝈√𝟐𝝅 −∞ √𝟐𝝅 −∞

𝒕𝟐 𝒕𝟐 𝒕𝟐
𝝈𝟐 ∞
− − 𝝁𝟐 − ∞ 2𝜇𝜎 ∞
 ∫ (𝑡 2 )𝒆 𝟐 𝑑𝑡 + √𝟐𝝅 ∫−∞ 𝒆 𝟐 𝑑𝑡 + √𝟐𝝅 ∫−∞ 𝒕𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞

(𝒙−𝝁)𝟐
𝟏∞ − 𝝈𝟐 𝝁𝟐 2𝜇𝜎2
 ∫ 𝒙𝟐 𝒆 𝟐𝝈𝟐
𝝈√𝟐𝝅 −∞
𝒅𝒙 =
√𝟐𝝅
𝑰𝟏 +
√𝟐𝝅
𝑰𝟐 + 𝑰
𝝈𝟐 √𝟐𝝅 𝟑
( Let us assume)

𝒕 𝟐
𝝈𝟐 ∞ 𝝈𝟐
𝟐 )𝒆− 𝟐
Let us Solve I1 = ∫ (𝒕 𝒅𝒕
√𝟐𝝅 √𝟐𝝅 −∞

𝒕𝟐
𝝈𝟐 ∞

 ∫ (𝑡 2 )𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞

𝒕𝟐
𝝈𝟐 ∞ −
 ∫ (𝑡) 𝑡𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞

 Integration by Parts

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


𝒕𝟐

 Put u = t and dv = 𝑡 𝒆 𝟐 𝑑𝑡

𝒕𝟐
 We get du =dt and v = ∫(𝒕)𝒆− 𝟐 𝒅𝒕

 We need to solve v now

𝒕𝟐 t2
2t
 Put 𝒆− 𝟐 =y and dy = e− 2 (− 2 ) dt = dy

t2
 So e− 2 (t)dt= - dy

𝒕𝟐 𝒕𝟐
 ∫(𝑡)𝒆− 𝟐 𝑑𝑡 = ∫ −𝑑𝑦 = -y =- 𝒆− 𝟐 = ……... (4)

𝒕𝟐 𝒕𝟐

∫(𝑡)𝒆 𝟐 𝑑𝑡 =- 𝒆− 𝟐 = -y

𝒕 𝟐
𝝈𝟐
∞ 2 )𝒆− 𝟐
 Now going back ∫ (𝑡 𝑑𝑡 = 𝑢𝑣 − ∫ 𝑣 𝑑𝑢
√𝟐𝝅 −∞

𝒕𝟐 𝒕𝟐
 Recalling u = t and dv = 𝑡 𝒆− 𝟐 𝑑𝑡 and v = =- 𝒆− 𝟐 (from (4))

𝒕𝟐
𝝈𝟐 − ∞
 From above results ∫ (𝑡 2 )𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞

𝒕𝟐 𝒕𝟐
𝝈𝟐 − −
 { t (- 𝒆 𝟐 ) - ∫−𝒆 𝟐 𝒅𝒕 }
√𝟐𝝅

 Now with limits

𝒕𝟐 𝟐
𝝈𝟐 − 𝝈𝟐 ∞ −𝒕
 t (- 𝒆 𝟐 ) |[ t= − ∞𝒕𝒐 𝒕 = ∞ ] +∫ ∫ 𝒆 𝟐 𝒅𝒕
√𝟐𝝅 √𝟐𝝅 −∞

𝒕𝟐
𝝈𝟐 ∞ −
 =𝟎+ + ∫ 𝒆 𝟐 𝒅𝒕
√𝟐𝝅 −∞

𝒕𝟐
𝟏 ∞
 𝝈𝟐 ( 𝟏√𝟐𝝅 ∫−∞ 𝒆− 𝟐 𝒅𝒕)

 𝝈𝟐 (using (3) with mean = 0 and sd = 1 area under the curve =


1 as it is a Standard Normal distribution )

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


𝝈𝟐
 I1 = 𝝈𝟐 ………………... (5)
𝝈√𝟐𝝅

𝒕 𝟐
𝝁𝟐 ∞
Let us Solve I2 = √𝟐𝝅 ∫−∞ 𝒆− 𝟐 𝑑𝑡

(𝒕−𝟎)𝟐 (𝒙−𝟎)𝟐
𝝁𝟐 ∞ − 𝝁𝟐 ∞ −
∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑡 = ∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑥 }
√𝟐𝝅 √𝟐𝝅

Using property of integral variables can be changed with same limits.

This is a standard normal distribution X ~𝑵 (𝝁, 𝝈𝟐 ) 𝑏𝑒𝑐𝑜𝑚𝑛𝑔 𝑿~𝑵(𝟎, 𝟏) with

(𝒙−𝟎)𝟐
𝝁𝟐 ∞ −
= ∫−∞
𝒆 𝟐(𝟏)𝟐 𝑑𝑥 )
√𝟐𝝅

Area under the curve for a standard normal distribution =1


Therefore 𝝁𝟐 (𝟏) = 𝝁𝟐 ……………………...(6)

∞ 𝒕𝟐
2𝜇𝜎 −
𝐋𝐞𝐭 𝐮𝐬 𝐒𝐨𝐥𝐯𝐞 𝑰𝟑 = ∫ 𝒕𝒆 𝟐 𝑑𝑡
√𝟐𝝅 −∞

𝒕𝟐 t2
2t
 Put 𝒆− 𝟐 =p and e− 2 (− ) dt = dp
2

 Change of limits: - Also, when t =−∞ 𝑎𝑛𝑑 𝑢 = 0 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑡 = ∞ , 𝑢 = 0

t2
 So e− 2 (t)dt= - dp

𝒕𝟐
2𝜇𝜎 −∞ 2𝜇𝜎 0
 ∫ (𝑡)𝒆 𝟐 𝑑𝑡 = √𝟐𝝅 ∫0 −𝑑𝑝 = 0 ……...(7)
√𝟐𝝅 −∞

(𝒙−𝝁)𝟐
𝟏 ∞ −
 Var (X) = ∫ 𝒙𝟐 𝒆
𝝈√𝟐𝝅 −∞
𝟐𝝈𝟐 𝒅𝒙 - 𝝁2

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


 Using (5), (6) and 7)

 𝜎 2 + 𝜇2 + 0 = 𝜇2

 𝜎 2 Proved
Var (x) = 𝜎 2

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Finding Median of the Normal Distribution
We know Median in Normal distribution is at x = 𝜇
We shall apply the method of Calculus to prove this
Let Me be the Median. We need to prove from the below Integral
that the Upper limit M = 𝝁

𝑴 𝟏
𝑾𝒆 𝒌𝒏𝒐𝒘 ∫−∞ 𝒇(𝒙)𝒅𝒙 =
𝟐
(𝒙−𝝁) 𝟐
𝑴 𝟏 − 𝟏
 ∫−∞ 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐 𝒅𝒙 =
𝟐

𝟐
𝟏 𝑴 −(𝒙−𝝁) 𝟏
 ∫ 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟐
𝝈√𝟐𝝅 −∞

 Using Integration by substitution

𝑥−𝜇
 Let = 𝑡 𝑎𝑛𝑑 𝑑𝑥 = 𝜎𝑑𝑡 substituting in the above integral
𝜎
𝑀−𝜇
Also when x =−∞ 𝑎𝑛𝑑 𝑡 = −∞ 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥 = 𝑀 , 𝑡 = 𝜎
𝑀−𝜇 𝒕𝟐
𝟏 − 𝟏
 ∫ 𝜎 𝒆 𝟐 𝜎𝑑𝑡 = 𝟐
𝝈√𝟐𝝅 −∞

𝑀−𝜇 𝒕𝟐
𝟏 − 𝟏
 ∫ 𝜎 𝒆 𝟐 𝑑𝑡 = 𝟐
√𝟐𝝅 −∞

𝑀−𝜇 (𝒕−𝟎)𝟐
𝟏 − 𝟏
 ∫ 𝜎
(𝟏)√𝟐𝝅 −∞
𝒆 𝟐(𝟏) 𝑑𝑡 = 𝟐

𝑀−𝜇 (𝒕−𝟎)𝟐
𝟏 − 𝟏
 ∫ 𝜎
(𝟏)√𝟐𝝅 −∞
𝒆 𝟐(𝟏) 𝑑𝑡 = 𝟐 is a Standard Normal Distribution with

𝜇 = 0 , 𝑀𝑒𝑎𝑛 𝑎𝑛𝑑 𝜎 = 1 , 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 whose Sum of Area =1

 If the Area = ½ and the Normal curve being symmetric, the Upper Limit

𝑀−𝜇
= 0 ,𝑀 − 𝜇 = 0 ,𝑀 = 𝜇
𝜎

 𝑃𝑟𝑜𝑣𝑒𝑑

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Finding Mode of the Normal Distribution
Mode is a random variable (x data) for which the data has the highest
probabilities Probability density function pdf f(x) is maximum
We know from the figure that Mode= Median = Mean = 𝑥 = 𝜇
for X ~𝑁 (𝜇, 𝜎 2 )
We know from the figure that Mode= Median = Mean = 𝑥 = 0
for Standard Normal Distribution X ~𝑁 (0,1)
Using the second derivative test, f(x) is maximum if the follow 2 conditions
are satisfied f’(x) = 0 and f’’(x) < 0
(𝒙−𝝁)𝟐
𝟏 −
Given f(x) = 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐

Differentiating with respect to x as 𝜇 𝑎𝑛𝑑 𝜎 are constants here


(𝒙−𝝁)𝟐
𝟏 − −𝟐(𝒙−𝝁)
 f ’(x) = 𝒆 𝟐𝝈𝟐 .
𝝈√𝟐𝝅 𝟐𝝈𝟐

(𝒙−𝝁)𝟐
𝟏 − −(𝒙−𝝁)
 f ’(x) = 𝒆 𝟐𝝈𝟐 .
𝝈√𝟐𝝅 𝝈𝟐

 Put f’(x) = 0

(𝒙−𝝁)𝟐
𝟏 − −(𝒙−𝝁)
 𝒆 𝟐𝝈𝟐 . =𝟎
𝝈√𝟐𝝅 𝝈𝟐

(𝑥−𝜇)2
− −(𝑥−𝜇)
 implies 𝑒 2𝜎2 = 0 𝑜𝑟 . =0
𝜎2

(𝑥−𝜇)2

 𝑒 2𝜎2 > 0 𝑎𝑙𝑤𝑎𝑦𝑠 𝑠𝑜 − (𝑥 − 𝜇) = 0 or x=𝜇

(𝒙−𝝁)𝟐 (𝒙−𝝁)𝟐
𝟏 − −𝟐(𝒙−𝝁) − 𝟏
 Now finding second derivative f’’ (x) , we get 𝝈√𝟐𝝅 { 𝒆 𝟐𝝈𝟐 . + 𝒆 𝟐𝝈𝟐 . (− 𝝈𝟐 )}
𝟐𝝈𝟐

(𝝁−𝝁)𝟐 (𝝁−𝝁)𝟐
𝟏 − −𝟐(𝝁−𝝁) − 𝟏
 Now f’’ (x=𝜇) = {𝒆 𝟐𝝈𝟐 . + 𝒆 𝟐𝝈𝟐 . (− 𝝈𝟐 )}
𝝈√𝟐𝝅 𝟐𝝈𝟐

𝟏 𝟏
 Now f’’ (x=𝜇) = { 𝒆𝟎 . (𝟎) + 𝒆𝟎 . (− 𝝈𝟐 )}
𝝈√𝟐𝝅

1
 As 𝜎 2 > 0 (− 𝜎2 ) < 0 𝑓𝑜𝑟 𝑎𝑛𝑦 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝜎 𝑠𝑜 𝑓 ′′ (𝜇) < 0 𝑓𝑜𝑟 𝑥 = 𝜇

 So f ’ (𝜇) 𝑎𝑛𝑑 𝑓 ′ (𝜇) < 0 means f (x is maximum at x= 𝜇

 Mode of f(x) is at x = 𝝁

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Interpreting Chebyshev’s Empirical rule of Normal distribution using
Calculus Methods

In Calculus methods, we need to prove


(𝒙−𝝁)𝟐
𝝁+𝝈 𝟏 −
1. ∫𝝁−𝝈 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟎. 𝟔𝟖

(𝒙−𝝁)𝟐
𝝁+𝟐𝝈 𝟏 −
2. ∫𝝁−𝟐𝝈 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟎. 𝟗𝟓
𝝈√𝟐𝝅

(𝒙−𝝁)𝟐
𝝁+𝟑𝝈 𝟏 −
3. ∫𝝁−𝟑𝝈 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟎. 𝟗𝟗𝟕
𝝈√𝟐𝝅

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Proof of (1)
Considering LHS

𝝁+𝝈
𝟏 (𝒙−𝝁)𝟐

∫ 𝒆 𝟐𝝈𝟐 𝒅𝒙
𝝁−𝝈 𝝈√𝟐𝝅

𝑥−𝜇
Let = 𝑡 𝑎𝑛𝑑 𝜎𝑑𝑡 = 𝑑𝑥 and dx =𝜎𝑑𝑡
𝜎

Substituting in the above integral

Change of Limits: - When x =𝝁 − 𝝈 𝑎𝑛𝑑 𝑡 = −1 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥 = 𝝁 + 𝝈 , 𝑡 = 1

𝒕𝟐
𝟏 1 −
 ∫ 𝒆 𝟐 (𝜎𝑑𝑡)
𝝈√𝟐𝝅 −1

1 𝒕𝟐
𝟏 −
 ∫ 𝒆 𝟐 (𝑑𝑡)
√𝟐𝝅 −1

 Proved
(𝒙−𝝁)𝟐
𝝁+𝝈 𝟏 −
∫𝝁−𝝈 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟎. 𝟔𝟖
𝝈√𝟐𝝅

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Proof of (2)
Considering LHS

𝝁+𝟐𝝈
𝟏 (𝒙−𝝁)𝟐

∫ 𝒆 𝟐𝝈𝟐 𝒅𝒙
𝝁−𝟐𝝈 𝝈√𝟐𝝅

𝑥−𝜇
Let = 𝑡 𝑎𝑛𝑑 𝜎𝑑𝑡 = 𝑑𝑥 and dx =𝜎𝑑𝑡
𝜎

Substituting in the above integral

Change of Limits: - When x =𝝁 − 𝟐𝝈 𝑎𝑛𝑑 𝑡 = −2 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥 = 𝝁 + 𝟐𝝈 , 𝑡 = 2

𝒕𝟐
𝟏 2 −
 ∫ 𝒆 𝟐 (𝜎𝑑𝑡)
𝝈√𝟐𝝅 −2

2 𝒕𝟐
𝟏 −
 ∫ 𝒆 𝟐 (𝑑𝑡)
√𝟐𝝅 −2

 Proved
(𝒙−𝝁)𝟐
𝝁+𝟐𝝈 𝟏 −
∫𝝁−𝟐𝝈 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟎. 𝟗𝟓𝟒𝟒 = 𝟎. 𝟗𝟓 𝒑𝒓𝒐𝒗𝒆𝒅

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com


Proof of (3)
Considering LHS

𝝁+𝟑𝝈
𝟏 (𝒙−𝝁)𝟐

∫ 𝒆 𝟐𝝈𝟐 𝒅𝒙
𝝁−𝟑𝝈 𝝈√𝟐𝝅

𝑥−𝜇
Let = 𝑡 𝑎𝑛𝑑 𝜎𝑑𝑡 = 𝑑𝑥 and dx =𝜎𝑑𝑡
𝜎

Substituting in the above integral

Change of Limits: - When x =𝝁 − 𝟑𝝈 𝑎𝑛𝑑 𝑡 = −3 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥 = 𝝁 + 𝟑𝝈 , 𝑡 = 3

𝒕𝟐
𝟏 3 −
 ∫ 𝒆 𝟐 (𝜎𝑑𝑡)
𝝈√𝟐𝝅 −3

3 𝒕𝟐
𝟏 −
 ∫ 𝒆 𝟐 (𝑑𝑡)
√𝟐𝝅 −3

 Proved
(𝒙−𝝁)𝟐
𝝁+𝟐𝝈 𝟏 −
∫𝝁−𝟐𝝈 𝝈√𝟐𝝅 𝒆 𝟐𝝈𝟐 𝒅𝒙 = 𝟎. 𝟗𝟗𝟕 = 𝟎. 𝟗𝟗𝟕 𝒑𝒓𝒐𝒗𝒆𝒅

Prof.Iyer CALCULUS IN PROBABILITY ib.mathfaculty@gmail.com

You might also like