OpenSAP Data Science W5

Week 5: Probability Distributions
Unit 1: Properties of Distributions

Properties of Distributions
Introduction
▪ A probability distribution is a mathematical function that

provides the probabilities of occurrence of different
possible outcomes in an experiment.
http://statisticsbyjim.com/basics/probability-distributions/
https://en.wikipedia.org/wiki/Probability_distribution
© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 2
Types of probability distribution
1. Discrete probability distribution
▪ The set of possible outcomes is discrete
2. Continuous probability distribution
▪ The set of possible outcomes can take on values in a

continuous range

Discrete probability functions
p(S)
1/6
0.16
0.14 5/36
0.12
1/9
0.10
0.08 1/12
0.06
1/18
0.04 The probability mass
1/36 function (pmf) of counts
0.02
from two dice
2 3 4 5 6 7 8 9 10 11 12
S
Discrete probability example
Number of
Probability
Heads
0 0.25
1 0.50
2 0.25
Flip a coin two times

Continuous probability functions
0.4
0.3
0.2
34.1% 34.1%
0.1
2.1% 2.1%
0.1% 13.6% 13.6% 0.1%
0.0
-3σ -2σ -2σ 0 1σ 2σ 3σ
The probability density function (pdf) of the normal distribution
Continuous probability example 1
a
Refer to https://stattrek.com/probability-distributions/discrete-continuous.aspx for more information.
Continuous probability example 2
http://onlinestatbook.com/2/calculators/normal_dist.html to draw a normal distribution

Continuous probability vs discrete probability distribution
Number of
Probability
Heads
0 0.25
1 0.50
2 0.25
Discrete Continuous

Summary
▪ A probability distribution is a mathematical function that

provides the probabilities of occurrence of different
possible outcomes in an experiment.
▪ A discrete random variable can take only a finite
number of different values like 0,1,2,3,4, etc., whereas
a continuous random variable is a variable that can
take an infinite number of possible values.
▪ Discrete probability functions are also known as
“probability mass functions” and can assume a
discrete number of values.
▪ Continuous probability functions are also known as
“probability density functions” and the probabilities
are measured over ranges of values rather than single
points.

Thank you.
Contact information:
open@sap.com
Follow all of SAP
www.sap.com/contactsap
© 2019 SAP SE or an SAP affiliate company. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of
SAP SE or an SAP affiliate company.
The information contained herein may be changed without prior notice. Some software products marketed by SAP SE and its
distributors contain proprietary software components of other software vendors. National product specifications may vary.
These materials are provided by SAP SE or an SAP affiliate company for informational purposes only, without representation or
warranty of any kind, and SAP or its affiliated companies shall not be liable for errors or omissions with respect to the materials.
The only warranties for SAP or SAP affiliate company products and services are those that are set forth in the express warranty
statements accompanying such products and services, if any. Nothing herein should be construed as constituting an additional
warranty.
In particular, SAP SE or its affiliated companies have no obligation to pursue any course of business outlined in this document or
any related presentation, or to develop or release any functionality mentioned therein. This document, or any related presentation,
and SAP SE’s or its affiliated companies’ strategy and possible future developments, products, and/or platforms, directions, and
functionality are all subject to change and may be changed by SAP SE or its affiliated companies at any time for any reason
without notice. The information in this document is not a commitment, promise, or legal obligation to deliver any material, code, or
functionality. All forward-looking statements are subject to various risks and uncertainties that could cause actual results to differ
materially from expectations. Readers are cautioned not to place undue reliance on these forward-looking statements, and they
should not be relied upon in making purchasing decisions.
SAP and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered
trademarks of SAP SE (or an SAP affiliate company) in Germany and other countries. All other product and service names
mentioned are the trademarks of their respective companies.
See www.sap.com/copyright for additional trademark information and notices.
Unit 2: The Normal Distribution
The Normal Distribution
Introduction
50
40
30
20
10
0
100 120 140 160 180 200
https://www.mathsisfun.com/data/standard-normal-distribution.html
https://en.wikipedia.org/wiki/Normal_distribution
Characteristics
The characteristics of the normal distribution: Mean

▪ mean = median = mode Median
▪ symmetry about the centre Mode
▪ 50% of values less than the mean and 50% greater
than the mean
Symmetry
50% 50%
Standard deviation
68% of values are

within 1 standard
deviation of the mean
95% of values are

within 2 standard
deviations of the mean
99.7% of values are

within 3 standard
For a standard deviation calculator, see:
https://www.mathsisfun.com/data/standard-
deviations of the mean
deviation-calculator.html
Standard normal distribution
Standardize
950 970 999 1010 1030 1050 1070 -3 -2 -1 0 +1 +2 +3
A Normal Distribution The Standard Normal Distribution
z is the "z-score" (standard score)

x is the value to be standardized
The formula for the z-score:
μ is the mean
σ is the standard deviation
For an interactive standard normal distribution calculator, see:

https://www.mathsisfun.com/data/standard-normal-distribution-table.html
Standard normal distribution example
19.1% 19.1%
15.0% 15.0%
0.5% 9.2% 9.2% 0.5%
4.4% 4.4%
0.1% 0.1%
1.7% 1.7%
Standard Normal Distribution

The empirical rule
99.7% of data is within 3 standard

deviations of the mean (തx − 3s to xത + 3s)
95% within 2 standard
deviations
68% within 1
standard deviation
0.4
0.3
0.2
34.1% 34.1%
0.1
2.1% 2.1%
0.1% 13.6% 13.6% 0.1%
0.0
𝑥−𝜇
-3σ -2σ -1σ 0 1σ 2σ 3σ 𝑧=
𝜎
ഥ
𝒙 − 𝟑𝒔 ഥ
𝒙 − 𝟐𝒔 ഥ
𝒙−𝒔 ഥ
𝒙 ഥ
𝒙+𝒔 ഥ
𝒙 + 𝟐𝒔 ഥ
𝒙 + 𝟑𝒔
Rules of thumb for detecting outliers
Possible Outliers Outliers

|z| > 2 |z| > 3
Not unusual
Moderately Moderately
unusual unusual
Outliers Outliers
Z = -3 Z = -2 Z = -1 Z=0 Z=1 Z=2 Z=3

Central limit theorem
population sampling distribution

distribution of the mean
https://en.wikipedia.org/wiki/Central_limit_theorem
https://machinelearningmastery.com/a-gentle-introduction-to-the-central-limit-theorem-for-machine-learning/
Summary
▪ The normal distribution is a very commonly encountered continuous probability

distribution.
▪ The characteristics of the normal distribution are:
− mean = median = mode
− symmetry about the centre
− 50% of values less than the mean and 50% greater than the mean
▪ When we calculate the standard deviation, we find that generally:
− 68% of values are within 1 standard deviation of the mean
− 95% of values are within 2 standard deviations of the mean
− 99.7% of values are within 3 standard deviations of the mean
▪ The empirical rule states that for a normal distribution, nearly all of the data will
fall within three standard deviations of the mean.
▪ The central limit theorem (CLT) establishes that when independent random
variables are added, their properly normalized sum tends towards a normal
distribution even if the original variables themselves are not normally distributed.
Thank you.
open@sap.com
Follow all of SAP

warranty.
Unit 3: Kurtosis and Skewness
Kurtosis and Skewness
Introduction to kurtosis
Positive Kurtosis
Negative Kurtosis
Normal Distribution
https://www.statisticshowto.datasciencecentral.com/probability-and-statistics/statistics-definitions/kurtosis-leptokurtic-platykurtic/
Kurtosis
▪ Data sets with high, positive kurtosis tend to ▪ Data sets with low kurtosis tend to have
have heavy tails, or outliers. light tails, or lack of outliers.
35
30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 2 3 4 5 6 7 8 9 10 11
▪ This distribution has positive kurtosis ▪ This distribution has low kurtosis (no tails)
(heavier tails compared to the normal
distribution)
https://en.wikipedia.org/wiki/Kurtosis
https://www.spcforexcel.com/knowledge/basic-statistics/are-skewness-and-kurtosis-useful-statistics
Excess kurtosis
0.8
D, 3
S, 2 Key:
L, 1.2
0.7 N, 0 Red, kurt 3, Laplace (D)ouble exponential
C, -0.59376
W, -1 distribution;
0.6 U, -1.2
Orange, kurt 2, hyperbolic (S)ecant distribution;
0.5 Green, kurt 1.2, (L)ogistic distribution;

Black, kurt 0, (N)ormal distribution;
0.4 Cyan, kurt −0.593762…, raised (C)osine
distribution;
0.3 Blue, kurt −1, (W)igner semicircle distribution;
Magenta, kurt −1.2, (U)niform distribution.
0.2
0.1
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
Kurtosis in financial markets
▪ Real estate (with a kurtosis of 8.75) and high yield

US bonds (8.63) are high risk investments.
▪ Investment grade US bonds (1.06) and small cap
US stocks (1.08) would be considered safer
investments.
Introduction to skewness
Left Normal: Right

(negative) Mean (positive)
skew: equals skew:
Mean mode Mean
is less than exceeds
mode mode
Value of function
Random variable
https://www.itl.nist.gov/div898/handbook/eda/section3/eda35b.htm
https://whatis.techtarget.com/definition/skewness
Mean and median
Skewed and Symetric Data

1 2 3 4 5 6
A B
8
6
4
Frequency 2
0
C
Data set A: Mean = 2.64 Median = 2.00
Data set B: Mean = 4.36 Median = 5.00
Data set C: Mean = 3.00 Median = 3.00
1 2 3 4 5 6
von Hippel, Paul T. (2005). "Mean, Median, and Skew: Correcting a Textbook Rule". Journal of Statistics Education. 13 (2).
https://en.wikipedia.org/wiki/Skewness
Why is skew important?
0,4
0,3
Positively Skewed Residuals Normal Distribution

Log
0,2
34,1% 34,1% Transformation

0,1
2,1% 2,1%
0,1% 13,6% 13,6% 0,1%
0,0
–30 –20 –10 μ 10 20 30

68.2%
Negatively Skewed Residuals Normal Distribution
95.4% Exponential
Transformation
99.7%
Normal Distribution Data Transformations
https://www.sheffield.ac.uk/polopoly_fs/1.579181!/file/stcp-marshallsamuels-NormalityS.pdf
https://www.quora.com/How-does-skewness-impact-regression-model
https://www.itl.nist.gov/div898/handbook/eda/section3/eda35b.htm
https://www.linkedin.com/pulse/question-does-skewness-variable-impact-predictive-data-mosaddar for more information
Summary
▪ Kurtosis is a measure of the "tailedness" of the probability distribution:

− Data sets with high kurtosis tend to have heavy tails, or outliers (“leptokurtic”).
− Data sets with low kurtosis tend to have light tails, or lack of outliers
(“platykurtic”).
− Distributions with zero excess kurtosis are called “mesokurtic”
(normal distribution family).
▪ Skewness is a measure of the asymmetry of a probability distribution.
− A distribution is symmetric if it looks the same to the left and right of the center
point.
− If most of the data is on the left side of the histogram but a few larger values are
on the right, the data is said to be skewed to the right (positive skew).
− If most of the data is on the right, with a few smaller values showing up on the left
side of the histogram, the data is skewed to the left (negative skew).
− If the distribution is symmetric, then the mean is equal to the median and the
distribution has zero skewness. If the distribution is both symmetric and
unimodal, then the mean = median = mode.
Thank you.
open@sap.com
Follow all of SAP

warranty.
Unit 4: Using the Normal Distribution to Calculate
Probability
Using the Normal Distribution to Calculate Probability
Normal distribution recap
Normal Curve Normal Curve

Smaller Standard Deviation Larger Standard Deviation

Probability and the normal distribution recap
a
Empirical rule recap
50% 50%
Total area under the The z values are the

curve = 100% number of std deviations
from the mean
34% 34%
2.35% 2.35%
0.15% 13.5% 13.5% 0.15%
μ-3σ μ -2σ μ -σ μ μ+σ μ+2σ μ+3σ
68%
95%
99.7%

Empirical rule example
Mean = 1.5
Question
▪ 95% of students at school are between
1.2m and 1.8m tall.
▪ Assuming this data is normally
distributed, calculate the mean and
standard deviation.
95%
2 SD 2 SD
http://davidmlane.com/hyperstat/z_table.html 4 SD

Find probabilities
▪ How can you use this theory in practice?

▪ To find the probability associated with a normal random variable,
use a graphing calculator, an online normal distribution calculator,
or a normal distribution table.
▪ There are lots of normal distribution calculators available online.
▪ Here are some examples for you:

https://www.mathportal.org/calculators/statistics-calculator/normal-distribution-calculator.php
https://stattrek.com/online-calculator/normal.aspx
https://www.hackmath.net/en/calculator/normal-distribution
http://davidmlane.com/hyperstat/z_table.html

Example 1
Question
▪ On average, a light bulb lasts 300
days with a standard deviation of
50 days.
▪ Assuming that bulb life is normally
distributed, what is the probability
that the light bulb will last at most
365 days?
https://www.hackmath.net/en/calculator/normal-distribution
https://www.hackmath.net/en/calculator/normal-distribution?mean=300&sd=50&above=&area=below&below=365&ll=&ul=&outsideLL=
&outsideUL=&draw=Calculate
Example 2
Question
▪ Scores on an IQ test are normally
distributed.
▪ If the test has a mean of 110 and a
standard deviation of 20, what is the
probability that a person who takes
the test will score between 90 and
120?
https://www.hackmath.net/en/calculator/normal-distribution?mean=110&sd=20&above=&below=&area=between&ll=90&ul=120&outsideLL=
Example 3
Question
▪ A student achieved a score
of 900 in an exam.
▪ The mean test score was
825 with a standard
deviation of 100.
▪ Assuming that test scores
are normally distributed,
what proportion of students
achieved a higher score
than 900?
https://www.hackmath.net/en/calculator/normal-distribution?mean=825&sd=100&area=above&above=900&below=&ll=&ul=&outsideLL=
Summary
▪ The normal distribution refers to a family of continuous probability

distributions.
▪ The area under the normal distribution curve can be used to calculate
probabilities for a normally distributed random variable.
▪ There are lots of normal distribution calculators available online.
Given the mean and standard deviation, the calculator can be used to
calculate the area under the normal curve (the probability):
− less than a value
− greater than a value
− between values
− outside two values
https://stattrek.com/probability-distributions/normal.aspx
https://www.mathsisfun.com/data/standard-normal-distribution.html
https://statistics.laerd.com/statistical-guides/normal-distribution-calculations.php
Thank you.
open@sap.com
Follow all of SAP

warranty.
Unit 5: Hypothesis Testing
Hypothesis Testing
Introduction

Hypothesis Testing
Null and alternative hypotheses
Determine whether a coin was fairly

balanced:
▪ A null hypothesis H0 might be that half the
flips would result in Heads and half in Tails.
▪ The alternative hypothesis Ha might be that
the number of Heads and Tails would be
very different.
Ho: P = 0.5
H1: P ≠ 0.5
α < 0.05
https://stattrek.com/hypothesis-test/hypothesis-testing.aspx
Hypothesis Testing
Testing
Hypothesis Testing
1. State the 2. Formulate an 3. Analyze 4. Interpret

hypotheses analysis plan sample data results

Hypothesis Testing
Decision rules
Region of Acceptance
Rejection Acceptance Rejection

Region Region Region
Reject H0 if test statistic is

in these critical regions
Hypothesis Testing
One-tailed and two-tailed tests
Two-Tailed Test
Non-Rejection Region
Reject Reject
Hypothesis Hypothesis
Left-Sided (One-Tailed) Test Right-Sided (One-Tailed) Test

Non-Rejection Region Non-Rejection Region
Reject Reject
Hypothesis Critical Values Hypothesis
α/2 = 0.5%
Critical Value α = 1% Critical Value α = 1%
http://www.stat.yale.edu/Courses/1997-98/101/sigtest.htm
https://blog.minitab.com/blog/adventures-in-statistics-2/understanding-hypothesis-tests-significance-levels-alpha-and-p-values-in-statistics
Hypothesis Testing
Decision errors
Truth
H0 is True H0 is False
Type II Error
H0 Not Rejected Correct Decision
β
Statistician’s opinion
(based on the sample
data and decision rule)
Type I Error
H0 Rejected Correct Decision
α
https://en.wikipedia.org/wiki/Power_(statistics)
Hypothesis Testing
Summary
▪ “Hypothesis testing” refers to the formal procedures used

by statisticians to accept or reject statistical hypotheses.
▪ There are two types of statistical hypotheses:
1. Null hypothesis (Ho) is usually the hypothesis that the
sample observations result purely from chance.
2. Alternative hypothesis (H1 or Ha) is the hypothesis that
the sample observations are influenced by some non-
random cause.
▪ An analysis plan includes decision rules for rejecting the null
hypothesis. Statisticians describe these decision rules in two
ways – with reference to a P-value or with reference to a
region of acceptance.
▪ Two types of errors can result from a hypothesis test.
− A Type I error occurs when the researcher rejects a null
hypothesis when it is true.
− A Type II error occurs when the researcher fails to reject a
null hypothesis that is false.

Thank you.
open@sap.com
Follow all of SAP

warranty.

OpenSAP Data Science W5

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

OpenSAP Data Science W5

Caricato da

Copyright:

Formati disponibili

Week 5: Probability Distributions

Unit 1: Properties of Distributions

▪ A probability distribution is a mathematical function that

1. Discrete probability distribution

▪ The set of possible outcomes is discrete

2. Continuous probability distribution

▪ The set of possible outcomes can take on values in a

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 3

Flip a coin two times

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 5

-3σ -2σ -2σ 0 1σ 2σ 3σ

The probability density function (pdf) of the normal distribution

http://onlinestatbook.com/2/calculators/normal_dist.html to draw a normal distribution

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 9

▪ A probability distribution is a mathematical function that

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 10

© 2019 SAP SE or an SAP affiliate company. All rights reserved.

The characteristics of the normal distribution: Mean

68% of values are

95% of values are

99.7% of values are

950 970 999 1010 1030 1050 1070 -3 -2 -1 0 +1 +2 +3

A Normal Distribution The Standard Normal Distribution

z is the "z-score" (standard score)

For an interactive standard normal distribution calculator, see:

Standard Normal Distribution

99.7% of data is within 3 standard

Possible Outliers Outliers

Z = -3 Z = -2 Z = -1 Z=0 Z=1 Z=2 Z=3

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 8

population sampling distribution

▪ The normal distribution is a very commonly encountered continuous probability

© 2019 SAP SE or an SAP affiliate company. All rights reserved.

0.5 Green, kurt 1.2, (L)ogistic distribution;

▪ Real estate (with a kurtosis of 8.75) and high yield

Left Normal: Right

Skewed and Symetric Data

Positively Skewed Residuals Normal Distribution

34,1% 34,1% Transformation

–30 –20 –10 μ 10 20 30

Normal Distribution Data Transformations

▪ Kurtosis is a measure of the "tailedness" of the probability distribution:

© 2019 SAP SE or an SAP affiliate company. All rights reserved.

Normal Curve Normal Curve

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 2

Total area under the The z values are the

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 4

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 5

▪ How can you use this theory in practice?

▪ Here are some examples for you:

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 6

▪ The normal distribution refers to a family of continuous probability

© 2019 SAP SE or an SAP affiliate company. All rights reserved.

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 2

Determine whether a coin was fairly

1. State the 2. Formulate an 3. Analyze 4. Interpret

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 4

Rejection Acceptance Rejection

Reject H0 if test statistic is

Left-Sided (One-Tailed) Test Right-Sided (One-Tailed) Test

Critical Value α = 1% Critical Value α = 1%

▪ “Hypothesis testing” refers to the formal procedures used

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 8

© 2019 SAP SE or an SAP affiliate company. All rights reserved.