Silver Oak College of Engineering and Technology

Subject Name: Probability, Statistics and Numerical Analysis

Faculty Name: Dr.Yogini Vashi
Unit -3 Statistics
Fitting of 𝒀 = 𝒂 + 𝒃𝒙
straight line Normal equations:

∑ 𝑦 = 𝑛𝑎 + 𝑏 ∑ 𝑥

∑ 𝑥𝑦 = 𝑎 ∑ 𝑥 + 𝑏 ∑ 𝑥 2

Fitting of (Second degree curve) 𝒚 = 𝒂 + 𝒃𝒙 + 𝒄𝒙𝟐

parabola
Normal equations:

∑ 𝑦 = 𝑛𝑎 + 𝑏 ∑ 𝑥 + 𝑐 ∑ 𝑥 2

∑ 𝑥𝑦 = 𝑛 ∑ 𝑥 + 𝑏 ∑ 𝑥 2 + 𝑐 ∑ 𝑥 3

∑ 𝑥2𝑦 = 𝑛 ∑ 𝑥2 + 𝑏 ∑ 𝑥3 + 𝑐 ∑ 𝑥4

Fitting of 𝒚 = 𝒂𝒃𝒙 or 𝒚 = 𝒂𝒙𝒃 or 𝒚 = 𝒂𝒆𝒃𝒙

exponential
curves (1) For the curve 𝒚 = 𝒂𝒃𝒙
Taking logarithm base e on both sides we get

log 𝑦 = log(𝑎𝑏 𝑥 )
log 𝑦 = log(𝑎) + log(𝑏 𝑥 )
log 𝑦 = log(𝑎) + 𝑥 log(𝑏)

where 𝑌 = log 𝑦 , 𝐴 = log 𝑎 , 𝑋 = 𝑥 , 𝐵 = log 𝑏

Now to fit the line 𝑌 = 𝐴 + 𝐵𝑋 use the normal equations

∑ 𝑌 = 𝑛𝐴 + 𝐵 ∑ 𝑋

∑ 𝑋𝑌 = 𝐴 ∑ 𝑋 + 𝐵 ∑ 𝑋 2
Solving above equations we get the value of A and B. Then find the
value of a and b from 𝑎 = 𝑒 𝐴 , 𝑏 = 𝑒 𝐵

(2) For the curve 𝒚 = 𝒂𝒙𝒃

Taking logarithm base e on both sides we get

log 𝑦 = log(𝑎𝑥 𝑏 )
log 𝑦 = log(𝑎) + log(𝑥 𝑏 )
log 𝑦 = log(𝑎) + 𝑏 log(𝑥)

where 𝑌 = log 𝑦 , 𝐴 = log 𝑎 , 𝑋 = log 𝑥 , 𝐵 = 𝑏

Now to fit the line 𝑌 = 𝐴 + 𝐵𝑋 use the normal equations

∑ 𝑌 = 𝑛𝐴 + 𝐵 ∑ 𝑋

∑ 𝑋𝑌 = 𝐴 ∑ 𝑋 + 𝐵 ∑ 𝑋 2

Solving above equations we get the value of A and B. Then find the
value of a and b from 𝑎 = 𝑒 𝐴 , 𝑏 = 𝐵

(3) For the curve 𝒚 = 𝒂𝒆𝒃𝒙

Taking logarithm base e on both sides we get

log 𝑦 = log(𝑎𝑒 𝑏𝑥 )
log 𝑦 = log(𝑎) + log(𝑒 𝑏𝑥 )
log 𝑦 = log(𝑎) + 𝑏𝑥 log(𝑒)

where 𝑌 = log 𝑦 , 𝐴 = log 𝑎 , 𝑋 = 𝑥 , 𝐵 = 𝑏

Now to fit the line 𝑌 = 𝐴 + 𝐵𝑋 use the normal equations

∑ 𝑌 = 𝑛𝐴 + 𝐵 ∑ 𝑋

∑ 𝑋𝑌 = 𝐴 ∑ 𝑋 + 𝐵 ∑ 𝑋 2

Solving above equations we get the value of A and B. Then find the
value of a and b from 𝑎 = 𝑒 𝐴 , 𝑏 = 𝐵
Mean It is denoted by symbol 𝑋̅

For ungrouped data

∑𝑥
𝑋̅ = 𝑛 𝑖
Where n represents the total number of observations

For grouped data

∑𝑓 𝑥
𝑋̅ = 𝑛𝑖 𝑖

Where 𝑛 = ∑ 𝑓𝑖
Median It is denoted by symbol M

For ungrouped data

𝑛+1 𝑡ℎ
𝑀=( ) observation (when n is an odd number)
2

1 𝑛 𝑡ℎ 𝑛+1 𝑡ℎ
𝑀 = 2 [( 2 ) obs + ( ) 𝑜𝑏𝑠] (when n is an even number)
2
Formula for grouped data

𝑛
−𝐹
𝑀 = 𝐿 + (2 )×𝑐
𝑓
where
= ∑ 𝑓𝑖

L=Lower limit of median class

F= Cumulative frequency of the class preceding the median class
f= Frequency of the median class
c= Class length of median class

Note: In case of continuous frequency distribution the class

corresponding to the cumulative frequency just greater than or equal to
𝑛
is called the median class.
2

Mode It is denoted by symbol Z

For ungrouped data:
An observation which occurs maximum number of times is called
mode value of that data.

Formula for grouped data

𝑓1 − 𝑓0
𝑍 =𝐿+( )×𝑐
2𝑓1 − 𝑓0 − 𝑓2
Where
 L= Lower limit of the modal class
𝑓0 = frequency of the class preceding the modal class
𝑓1 = frequency of the modal class
𝑓2 = frequency of the class succeeding the modal class
𝑐 = class length of the modal class

Relation between mean median and mode

̅
𝒁 = 𝟑𝑴 − 𝟐𝑿
Standard
deviation It is denoted by the SD or symbol 𝜎 (𝑠𝑖𝑔𝑚𝑎)

For ungrouped data

∑(𝑥𝑖 −𝑥̅ )2 ∑ 𝑥𝑖 2
𝜎=√ OR 𝜎=√ − 𝑥̅ 2
𝑛 𝑛

For grouped data

∑ 𝑓𝑖 (𝑥𝑖 −𝑥̅ )2 ∑ 𝑓𝑖 𝑥𝑖 2
𝜎=√ OR 𝜎=√ − 𝑥̅ 2
𝑛 𝑛

Variance = 𝝈𝟐

Coefficient of It is denoted by symbol CV

Variation
𝜎
CV = × 100
𝑥̅

Moments about mean (central moments)

Moments For ungrouped data

∑𝑛
𝑖=1(𝑥𝑖 −𝑥̅ )
𝑟
𝜇𝑟 = 𝑟 = 1,2,3,4
𝑛

For grouped data

∑𝑛
𝑖=1 𝑓𝑖 (𝑥𝑖 −𝑥̅ )
𝑟
𝜇𝑟 = 𝑟 = 1,2,3,4
𝑛

Moments about assume mean (raw moments):

In above four formulae if we take 𝑥̅ = 𝑎 , we get formula for raw

moments. Where a is assume mean or arbitrary origin.
 ∑𝑛
𝑖=1(𝑥𝑖 −𝑎)
𝑟
𝜇𝑟 ′ = for 𝑟 = 1,2,3,4
𝑛

∑𝑛
𝑖=1 𝑓𝑖 (𝑥𝑖 −𝑥̅ )
𝑟
𝜇𝑟 ′ = (for grouped data)
𝑛

Moments about zero (simple moments):

In the central moments formula if we take 𝑥̅ = 0 then we get the formula
for simple moments.

Relation between central moments and raw moments:

𝜇1 = 𝜇1 ′ − 𝜇1 ′

𝜇2 = 𝜇2 ′ − (𝜇1 ′ )2

𝜇3 = 𝜇3 ′ − 3𝜇2 ′ 𝜇1 ′ + 2(𝜇1 ′ )3

𝜇4 = 𝜇4 ′ − 4𝜇3 ′ 𝜇1 ′ +6𝜇2 ′ (𝜇1 ′ )2 − 3(𝜇1 ′ )4

Relation between raw moments and central moments:

𝜇1 ′ = 𝜇 − 𝑎

𝜇2 ′ = 𝜇2 + (𝜇1 ′ )2

𝜇3 ′ = 𝜇3 + 3𝜇2 ′ 𝜇1 ′ + 2(𝜇1 ′ )3

𝜇4 ′ = 𝜇4 + 4𝜇3 𝜇1 ′ + 6𝜇2 (𝜇1 ′ )2 + (𝜇1 ′ )4

Skewness:
𝜇3 2
𝛽1 = 3
𝜇2

Karl Pearson’s coefficients of skewness:

𝑀𝑒𝑎𝑛 − 𝑚𝑜𝑑𝑒
𝑆𝑘 =
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛

Kurtosis:

𝜇4
𝛽2 =
𝜇2 2

Properties of central moments:

(a) The first moment about mean is always zero. (𝜇1 = 0)
(b) The second moments about mean is variance. (𝜇2 = 𝜎 2 )
 (c) In a symmetric distribution all odd moments are zero.(𝜇1 = 𝜇3 =
𝜇2𝑟+1 = 0)

Unit-4 Correlation and regression

1 Correlation coefficient is denoted by symbol r (−𝟏 ≤ 𝒓 ≤ 𝟏):

∑(𝑥−𝑥̅ )(𝑦−𝑦̅)
𝑟=
√∑(𝑥−𝑥̅ )𝟐 √∑(𝑦−𝑦̅)𝟐

OR
∑𝑥 ∑𝑦
∑ 𝑥𝑦−
𝑛
𝑟= 2 2
√∑ 𝑥 2 −(∑ 𝑥) √∑ 𝑦 2 −(∑ 𝑦)
𝑛 𝑛

OR

∑ 𝑑𝑥 ∑ 𝑑𝑦
∑ 𝑑𝑥 𝑑𝑦 −
𝑛
𝑟=
2
2
√∑ 𝑑 2 − (∑ 𝑑𝑦 ) √∑ 𝑑𝑥 2 − (∑ 𝑑𝑥 )
𝑦 𝑛 𝑛

Where 𝑑𝑥 = 𝑥 − 𝑎 , 𝑑𝑦 = 𝑦 − 𝑏
a and b are arbitrary origin.
Spearman’s rank correlation coefficient:

6 ∑ 𝑑2
𝑟 =1−
𝑛(𝑛2 − 1)
where ∑ 𝑑 = ∑(rank of 𝑥 − rank of 𝑦)

Formula of Spearman’s rank correlation coefficient for tied rank:

1 1
6 [∑ 𝑑 2 + 12 (𝑚1 3 − 𝑚1 ) + 12 (𝑚2 3 − 𝑚2 )+. . .]
𝑟 =1−
𝑛(𝑛2 − 1)

where m is the number of items having equal rank.

2 Regression:

Line of regression of x on y

𝑥 − 𝑥̅ = 𝑏𝑥𝑦 (𝑦 − 𝑦̅)

where
 ∑(𝑥−𝑥̅ )(𝑦−𝑦̅)
𝑏𝑥𝑦 = ∑(𝑦−𝑦̅)𝟐
(Regression coefficient of x on y)

OR

∑𝑥 ∑𝑦
∑ 𝑥𝑦−
𝑛
𝑏𝑥𝑦 =
(∑ 𝑦)2
√∑ 𝑦 2 −
𝑛

OR

∑ 𝑑𝑥 ∑ 𝑑𝑦
∑ 𝑑𝑥 𝑑𝑦 −
𝑛
𝑏𝑥𝑦 =
2
√∑ 𝑑 2 −(∑ 𝑑𝑦 )
𝑦 𝑛

Line of regression of y on x

𝑦 − 𝑦̅ = 𝑏𝑦𝑥 (𝑥 − 𝑥̅ )

where

∑(𝑥−𝑥̅ )(𝑦−𝑦̅)
𝑏𝑦𝑥 = ∑(𝑥−𝑥̅ )𝟐
(Regression coefficient of y on x)

OR

∑𝑥 ∑𝑦
∑ 𝑥𝑦−
𝑛
𝑏𝑦𝑥 =
(∑ 𝑥)2
√∑ 𝑥 2 −
𝑛

OR

∑ 𝑑𝑥 ∑ 𝑑𝑦
∑ 𝑑𝑥 𝑑𝑦 −
𝑛
𝑏𝑦𝑥 = 2
√∑ 𝑑𝑥 2 −(∑ 𝑑𝑥 )
𝑛

Correlation coefficient 𝑟 = √𝑏𝑥𝑦 𝑏𝑦𝑥