Sie sind auf Seite 1von 17

AB1202

Statistics and Analysis


Lecture 1 (Part 1 of 2)
Concepts of Probability
Chin Chee Kai
cheekai@ntu.edu.sg
Nanyang Business School
Nanyang Technological University
NBS 2016S1 AB1202 CCK-STAT-018
2

Concepts of Probability
• Probability and Counting
• Elementary Probability Rules
• Conditional Probability
• Joint, Marginal Probabilities
• Permutations and Combinations
• Independent and Mutually Exclusive Events
• Populations and Samples
• Sample Spaces and Events
• Data and Grouped Data
NBS 2016S1 AB1202 CCK-STAT-018
3

Probability and Counting


𝐶𝑜𝑢𝑛𝑡 𝑜𝑓 𝑤𝑎𝑛𝑡𝑒𝑑 𝑖𝑡𝑒𝑚𝑠 𝐶 𝑊
• Probability = =
𝐶𝑜𝑢𝑛𝑡 𝑜𝑓 𝐴𝐿𝐿 𝑖𝑡𝑒𝑚𝑠 𝐶(𝑈)
where we use W=set of wanted items, and U=set of
ALL items
𝐶 𝐴 5
• Probability of A=𝑃 𝐴 = =
𝐶(𝑈) 10
𝐶 𝐵 3
• Probability of B= 𝑃 𝐵 = =
𝐶(𝑈) 10

A B
U
NBS 2016S1 AB1202 CCK-STAT-018
4

Elementary Probability Rules


𝐶 𝐴′ 𝐶 𝑈 −𝐶 𝐴
• 𝑃 𝐴′ = = =1−𝑃 𝐴
𝐶 𝑈 𝐶 𝑈
𝐶 𝐴∪𝐵 𝐶 𝐴 +𝐶 𝐵 −𝐶(𝐴∩𝐵)
• 𝑃 𝐴∪𝐵 = =
𝐶 𝑈 𝐶 𝑈
=𝑃 𝐴 +𝑃 𝐵 −𝑃 𝐴∩𝐵
= 1 − 𝑃(𝐴′ ∩ 𝐵′ )

𝐶 𝑈 10
• Clearly, 𝑃 𝑈 = = =1
𝐶(𝑈) 10
• Clearly, 𝑃 𝑊 ≥ 0
for any subset 𝑊 ∈ 𝑈
A B
U
NBS 2016S1 AB1202 CCK-STAT-018
5

Conditional Probability
𝐶 𝐴∩𝐵 𝐶 𝐴∩𝐵 /𝐶(𝑈) 𝑃 𝐴∩𝐵
• 𝑃 𝐴|𝐵 = = =
𝐶 𝐵 𝐶 𝐵 /𝐶(𝑈) 𝑃 𝐵
𝐶 𝐵∩𝐴 𝐶 𝐵∩𝐴 /𝐶(𝑈) 𝑃 𝐵∩𝐴
•𝑃 𝐵𝐴 = = =
𝐶 𝐴 𝐶 𝐴 /𝐶(𝑈) 𝑃 𝐴

• Clearly, 𝑃 𝐴|𝐵 𝑃 𝐵 = 𝑃 𝐵 𝐴 𝑃 𝐴 = 𝑃(𝐴 ∩ 𝐵)

𝐶 𝐴∩𝐵 2
• Eg: 𝑃 𝐴|𝐵 = =
𝐶 𝐵 3
𝐶 𝐵∩𝐴 2
• Eg: 𝑃 𝐵|𝐴 = =
𝐶 𝐴 5

A B
U
NBS 2016S1 AB1202 CCK-STAT-018
6

Joint, Marginal Probabilities


Joint of X & Y: P(X=x, Y=y) Marginal: P(X=x), P(Y=y)

Means the probability of X Means consider the


being of value x AND Y being probability of X being of value
of value y simultaneously. x, with Y ranging over all
possible events of Y.
Separately, consider the
probability of Y being of value
y, with X ranging over all
X
possible events of X.
X Y
5 10
5 10 100 200
100 0.3 0.4 0.7
Y 0.4 0.6 0.7 0.3
200 0.1 0.2 0.3
0.4 0.6
This is called a “contingency table”
NBS 2016S1 AB1202 CCK-STAT-018
7

Joint, Marginal and Conditional


Probabilities
• Given a joint probability table, find:
▫ P(X=5 | Y=100)
 Means find probability of X=5 given that Y is 100.
𝑃 𝑋=5 ∩ 𝑌=100 0.3 X
 Means = = 0.4286
𝑃 𝑌=100 0.7
5 10
▫ P(Y=200 | X=10)
100 0.3 0.4 0.7
 Means find probability of Y=200 Y
given that X is 10. 200 0.1 0.2 0.3
𝑃 𝑌=200 ∩ 𝑋=10 0.2
 Means = = 0.3333. 0.4 0.6
𝑃 𝑋=10 0.6

▫ P(X=10 | Y=100 or X=5)  Can you find this?


▫ Answer is in invisible ink as follows:
𝑃 𝑋=10 ∩(𝑌=100 ∪ 𝑋=5) 0.4
▫ = = 0.5
𝑃 𝑌=100 ∪ 𝑋=5 0.4+0.3+0.1
NBS 2016S1 AB1202 CCK-STAT-018
8

Joint, Marginal and Conditional


Probabilities
• Given a joint probability table, find:
▫ P(Has Visa | Has Master)
 Means find probability of having Visa given that person has
Master.
𝑃 𝑉𝑖𝑠𝑎 ∩ 𝑀𝑎𝑠𝑡𝑒𝑟 0.3
 Means = = 0.4286 X
𝑃 𝑀𝑎𝑠𝑡𝑒𝑟 0.7

▫ P(No Master | No Visa) Visa No


Visa
 Means find probability of having no
Master 0.3 0.4 0.7
Master given that person has no Visa. Y
𝑃 𝑁𝑜 𝑀𝑎𝑠𝑡𝑒𝑟 ∩ 𝑁𝑜 𝑉𝑖𝑠𝑎 0.2 No 0.1 0.2 0.3
 Means = = Master
𝑃 𝑁𝑜 𝑉𝑖𝑠𝑎 0.6
0.3333. 0.4 0.6

▫ P(No Visa | Has Master or Visa)  Can you find this?


▫ Answer is same as invisible ink in previous slide
NBS 2016S1 AB1202 CCK-STAT-018
9

Permutations and Combinations


• Let’s count quickly, and accurately.
• If 2 balls are picked in sequence, how many sequences
are there to produce a sum of 3? Answer: 2
• How many possible outcomes are there to pick 2 balls
9!
in sequence? Answer: 9 × 8 = 9𝑃2 = = 72
9−2 !
• What is probability of picking 2 balls such that their
2
sum is 3? = 0.0278
72
• What is probability of picking 2 balls such that
4
product is 24? = 0.0556
72

1 2 3 4 5 6 7 8 9
NBS 2016S1 AB1202 CCK-STAT-018
10

Permutations and Combinations


• If we scoop 2 balls at once, how many different
pairs would add up to 3? Answer: 1
• How many possible pairs are there with a scoop of
9×8 9 9!
2 balls? Answer: = 𝐶2 = = 36
2 9−2 !2!
• What is probability of scooping 2 balls such that
1
their sum is 3? = 0.0278 (same as before)
36
• What is probability of scooping 2 balls such that
2
product is 24? = 0.0556
36

1 2 3 4 5 6 7 8 9
NBS 2016S1 AB1202 CCK-STAT-018
11

Independent Events
• An event, A, that does not enhance or reduce the
probability of occurrence of another event, B, is
physically independent from event B.
• Two events, A and B, which satisfy the following:
𝑃 𝐴 ∩ 𝐵 = 𝑃 𝐴 𝑃 𝐵 ⟺ 𝑃 𝐵 = 𝑃(𝐵|𝐴)
is said to be statistically independent.
Physical independence ⟹ statistical independence
• A random variable, X, is Is event A independent
independent from another of itself?
random variable, Y, if the
occurrence of one does not
change the distribution of the
other.
• 𝑃 𝑋 = 𝑥, 𝑌 = 𝑦 = 𝑃 𝑋 = 𝑥 𝑃 𝑌 = 𝑦
A B
U
NBS 2016S1 AB1202 CCK-STAT-018
12

Mutually Exclusive Events


• If A, B are mutually exclusive events, then they are
NOT independent.
• 𝑃 𝐴∩𝐵 =0
4 2
•𝑃 𝐴 𝑃 𝐵 = ×
• 𝑃 𝐴∩𝐵 ≠𝑃 𝐴 𝑃 𝐵
10 10
A B
U

• But if A, B are NOT independent, they don’t


necessarily have to be mutually exclusive.
• And if A, B are independent, they don’t necessarily
have to be mutually exclusive either.
▫ This is clear from previous example in the diagram.
NBS 2016S1 AB1202 CCK-STAT-018
13

Populations and Samples


• Population: set of all interested values
• Can be finite or infinite
▫ Eg: Finite: quiz scores of a class of AB1202 students
▫ Eg: Infinite: quiz scores of all (past and future)
AB1202 students
• “Interested values” could be any data we care to
study and measure.
▫ Eg: quiz score, height, body temperature, favorite
subjects, etc
NBS 2016S1 AB1202 CCK-STAT-018
14

Populations and Samples


• Sample: subset of a predefined population
• Why do we take a sample from the population?
▫ Population is just too large (expensive, time-
consuming, exhausting, difficult, etc) to study
• Sample size, n, is always finite
• How should we sample from population?
▫ Can we always pick the same 5 items to measure?
▫ Can we just pick the 5 most convenient items?
▫ Can we choose what we like best to measure?
▫ Can we save cost by measuring only 1 item?
▫ Most importantly: Samples must be randomly
chosen
NBS 2016S1 AB1202 CCK-STAT-018
15

Samples Spaces and Events


• An outcome is the unique measurement value (or
observed state).
▫ Eg: 1 to 6 is each an outcome of throwing a dice.
• The set of all possible unique outcomes of an
intended measurement (or study, or statistical
experiment) is the sample space.
▫ Eg: Sample space of throwing dice: {1,2,3,4,5,6}
• An event is a collection of outcomes (ie, a subset)
within the sample space.
▫ Eg: Event of odd values: {1,3,5}
𝐶 𝑂𝑑𝑑 𝑣𝑎𝑙𝑢𝑒 𝐸𝑣𝑒𝑛𝑡 3
• Probability of odd values: = = 0.5
𝐶 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑝𝑎𝑐𝑒 6

• Notice this is another concept of Probability


NBS 2016S1 AB1202 CCK-STAT-018
16

Data and Grouped Data


• Measured data are recorded as:
▫ 5, 6, 5, 5, 4, 4, 6, 5, 7, 4 (N=10)
• Sum of Data= 𝑥1 + 𝑥2 + ⋯ + 𝑥10 = 𝑁 𝑖=1 𝑥𝑖
• This may be represented also with unique values:
5
Frequency
𝑥′𝒊 4 5 6 7 4
3

Frequency 𝑓𝑖 3 4 2 1 2
1
0

• This is called “Grouped Data”. 4 5 6 7

4
• Sum of Data= 𝑥′1 ∙ 𝑓1 + ⋯ + 𝑥′4 ∙ 𝑓4 = 𝑖=1 𝑥′𝑖 ∙ 𝑓𝑖
• Advantages are (i) summarized, (ii) chart-ready, (iii) shows
mode readily, (iv) can show tons of data in limited space
• You will see grouped data a lot, both in this course and in
practice.
NBS 2016S1 AB1202 CCK-STAT-018
17

Common Series To Take Note


𝑘 𝑥2 𝑥3
∞ 𝑥
• 𝑒𝑥 = 𝑘=0 𝑘! =1 + 𝑥 + + +⋯
2! 3!

∞ −1 𝑘+1 𝑥 𝑘 𝑥2 𝑥3 𝑥4
• ln 1 + 𝑥 = 𝑘=1 =𝑥 − + − +⋯
𝑘 2 3 4
for 𝑥 ∈ (−1,1]
1 ∞ 𝑘
• = 𝑘=0 𝑥 = 1 + 𝑥 + 𝑥2 + 𝑥3 + ⋯ for 𝑥 ∈ (−1,1)
1−𝑥

1 𝑛
• 𝑛 𝑛+1 = 𝑘=1 𝑘 = 1 +2 + 3 + ⋯+ 𝑛
2

1 𝑛 2
• 𝑛 𝑛 + 1 (2𝑛 + 1) = 𝑘=1 𝑘 = 12 + 22 + 32 + ⋯ + 𝑛2
6

Das könnte Ihnen auch gefallen