Skewness-Kurtosis

and
EDA
Dr.Srilakshminarayana.G

Skewness and Kurtosis

Skewness
Measure of the degree of asymmetry of a frequency distribution
Skewed to left
Symmetric or unskewed
Skewed to right

Kurtosis
Measure of flatness or peakedness of a frequency distribution
Platykurtic (relatively flat)
Mesokurtic (normal)
Leptokurtic (relatively peaked)

Negative Skewness
Skewed to left

Symmetric
Symmetric

Positive Skewness
Skewed to right

Symmetric Bimodal Distribution

Symmetric distribution with two Modes
Mean = Median
40
35

35

Frequency

30

20

20
15

15

10

10

10

0
100

200

300

400
X

500

600

700

Kurtosis
Platykurtic - flat distribution

Kurtosis
Mesokurtic - not too flat and not too peaked

Kurtosis
Platykurtic - flat distribution

Kurtosis
Mesokurtic - not too flat and not too peaked

Kurtosis
Leptokurtic - peaked distribution

Relations between the Mean and Standard

Deviation
Chebyshevs Theorem
Applies to any distribution, regardless of shape
Places lower limits on the percentages of observations
within a given number of standard deviations from the
mean

Empirical Rule
Applies only to roughly mound-shaped and symmetric
distributions
Specifies approximate percentages of observations within a
given number of standard deviations from the mean

Chebyshevs Theorem

At least
of the elements of any distribution
k 2
lie within k standard deviations of the mean
1

At
least

1
1 3
=
1

= = 75%
2
4
4
2

1
1 8
1 2 = 1 = = 89%
9 9
3
1

1
1 15
=
1

=
= 94%
2
16
16
4

2
Lie
within

3
4

Standard
deviations
of the mean

Empirical Rule
For roughly mound-shaped and symmetric
distributions, approximately:
68%

95%

All

1 standard deviation
of the mean
Lie
within

2 standard deviations
of the mean
3 standard deviations
of the mean