lOMoARcPSD|6274171

Full summary of 2DI90 from book and lectures

Probability and statistics (Technische Universiteit Eindhoven)

Studocu is not sponsored or endorsed by any college or university

Downloaded by ciko miko ([email protected])
 lOMoARcPSD|6274171

Week 1
Video 1: MR 2.1: Introduction+Motivation, Sample spaces and events
Probabilistic models

• Simplifications

Random experiment

• An experiment that can result in different outcomes, even though it is repeated in the
same manner every time

Sample spaces

• The set of all possible outcomes of a random experiment

o the sample space of an experiment
▪ denoted as S

discrete sample space

• If it consist of a finite or countable infinite set of outcomes

Continuous sample space

• If it contains an interval (either finite or infinite) of real numbers

Event

• A subset of the sample space of random experiment

Union

• All events that are in E1 or E2

Intersection

• All events that are in E1 and E2

Complement

• All events that are not in E1

Mutually exclusive

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

• Two events have nothing the in common

Video 2: MR 2.2 Counting techniques and Permutations

Counting techniques

• Decision tree / tree diagram

Permutations

• The number of permutations of n different elements is n!

o

Permutations of subsets

• The number of permutations of subsets of r elements selected from a set of n

different elements is

Combinations

• The number of combinations, subsets of r elements that can be selected from a set
of n elements, is denoted as

Replacement

• Placing the element back after sampling

o Changes the formula

Video 3: MR 2.3 Interpretations and Axioms of Probability, MR 2.4 Addition Rules

• Frequentist approach / probability
o Limit of its relative frequency in many trials

• Probability of an event

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

o For a discrete sample space, denoted as P(E), equals the sum of the
probabilities of the outcomes in E

• Equally likely outcomes( Laplace experiment)

o Whenever a sample space consists of N possible outcomes that are equally
likely, the probability of each outcome is 1/N
o This kind of experiment is called Laplace experiment

• Axioms of probability

• Conclusions

Probability of a union

Video 4: MR 2.5 Conditional probability

Conditional probability

• The probability of an event B under the knowledge that the outcome will be in event
A is denoted as P(B|A)
o This is called the conditional probability of B given A

Video 5: MR2.6 Multiplication Rules; MR2.7 Independence; MR2.8 Bayes' Theorem

Multiplication rule

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Total probability rule

•
• Total probability rule (multiple events)

Independence

• Two events are independent if any one of the following equivalent statements is true

o
o Multiple events

▪
o Lemma

Bayes’ Theorem

•
• Multiple events

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Independences formula

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 2
Video 2: MR2.9-3.2
Random variable

• A random variable is a function that assigns a real number to each outcome in the
sample space of a random experiment
• Notation
o Random variable
▪ Uppercase letter
o Observed value of random variable
▪ Lowercase letter

Discrete random variable

• A discrete random variable is a random variable with a finite (or countably infinite)
range

Probability mass function

•
• Example

Cumulative distribution function

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 2: MR3.3
Mean or expected value of the discrete random variable X for discrete random variables

Variance of X

Standard deviation of X is given by

Expected value of a function of a random variable

Special cases

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 3: MR3.4-3.6
Discrete uniform distribution

•
• Mean and variance

Bernoulli distribution

•
• Mean and variance

Geometric distribution

Binomial distribution

• Item drawn is put back

o With replacement

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

•
• Mean and variance

Video 4: MR 3.6-MR3.8
Negative binomial distribution

•
• Mean and variance

Hypergeometric distribution

• Don’t put the item drawn back

o Without replacement

•
• Mean and variance

Poisson distribution

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Poisson random variable

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 3
Video 1: MR4.1-MR4.3
P(X = a) = 0 for all a in R

• 0 probability to draw a specific number

Probability density function

Important note continuous random variables:

•
• 4.1

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Continues vs. discrete random variables

Cumulative distribution function

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Mean and variance continuous random variable

Video 2: MR4.3 – MR 4.5

Expected value of a function of a random variable

Properties of mean and variance

• Same as in discrete case

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Uniform continuous distribution:

Normal distribution

• Also called Gaussian distribution

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Standard normal distribution

Standard normal distribution

Normal approximation to the binomial distribution

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Normal approximation to the poisson distribution don’t forget the correction!!!

Video 3: MR4.7
Exponential distribution

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 4: MR4.8 + 4.9

Erlang distribution

Gamma distribution

•
• Not have to be done by hand
• Watch out for the documentation used

Weibull distribution

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

•
•

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 4
Video 1: MR5.1-MR5.2
Joint probability mass function (discrete random variables)

•
𝑝3
(𝑝1 𝑝2
𝑥 )×( 𝑦 )×(𝑘−𝑥−𝑦)
o 𝑓𝑥𝑦(𝑥, 𝑦) =
(𝑛
𝑘)
▪ n = full group
▪ k = sample size
▪ P1, p2 and p3 people in condition 1,2 and 3 compared to full group
• Can be calculated from percentages and n
▪ X people of k in condition of p1
▪ Y people of k in condition of p2

27b66

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Marginal probability mass function (discrete variables)

Conditional probability mass function (discrete random variable)

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Independence of random variables

(conditional mean and variance)

Video 2: MR 5.3 – MR5.4

Mean of a function of two random variables

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Covariance and correlation

•
• Example:

o
o E(XY):

o
•

•
o If they are 0, it does not have to that they are independent

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 3: MR 5.6
Sum of random variables

Linear functions of normal distributions

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Law of large numbers

Central limit theorem (CLT)

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Homework:
Multinomial distribution

Examples:

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 5
Video 1: MR6.1
Sample mean:

Sample variance:

Sample range:

Order statistics:

•
• Sorting the data

Sample median:

Sample percentiles

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 2: MR6.3-6.7

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 6
Video 1: MR7.1 – MR7.2
Sampling distribution

• Random sample

o
• Statistic and sampling distribution

Estimator and estimate `

Central limit theorem (CLT)

video 2: MR7.3
unbiased estimator

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Variance and standard error of an estimator

Mean squared error of an estimator / bias-variance decomposition

Relative efficiency

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

•
Standard error of sample mean:

Estimator for mean:

Estimator mean and variance:

Sum of squares

example

•
o Finding out if unbiased:

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

o Finding variance of estimator

Video 3: MR7.4_part1
Method of moments

Maximum likelihood estimation

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 4: MR7.4_part2
Maximum likelihood equation

Properties of the MLE

Video 5: MR8.1
Interval estimate and confidence interval

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Confidence interval on the mean of a normal distribution, variance known

General method to derive a confidence interval

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 7
Video 1: MR8.2-8.4
Student-distribution or t-distribution

Confidence interval on the mean of a normal distribution, variance unknown

Confidence interval summary:

Chi square distribution

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Confidence intervals on the variance

•
o If you do it with variance, then do everything with variance and square root
the final intervals
Confidence interval for a population proportion

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 2: MR8.7.1
Prediction interval for a future observation

Video 3: MR9.1_Part1
Test statistic

Statistical test and critical Region

Video 4: MR9.1_Part2
p-value:

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

• Z0-value =
• Onesided p test

o
• Two sided p test

Video 5: MR 9.1 – 9.3

General procedure for hypothesis testing

Tests on the mean of a normal distribution, variance known

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Probability of a type 2 error variance known

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

•
o u = true mean
o u0 = null hypothesis mean
• Power = 1-B

Choice of sample size (two sided test)

Choice of sample size (one sided test)

Tests on the mean of a normal distribution, variance unknown

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Type I error

Type II error

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Week 8
Video 1: MR9.4_MR9.5
Chi square distribution

•
Binomial proportion hypothesis testing

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Type 2 error binomial proportion hypothesis testing

•
Video 2: MR9.7
Goodness of fit (peason Chi-squared test)

•
o E is calculated using the formula corresponding to the distribution type per
value
▪ Then using this percentage to multiply with n
o O equals the number observed frequencies
o Choose the chi squared alpha value with n = the number of bins
▪ So fill in n – 1 = number of bins - 1

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

•
o So change the binomial with the type in the question

Video 3: MR11.1_MR11.2
Simple linear regression

Method of least squares

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

• = sum of yi / n

• = sum of xi / n

Sum of squared error

Residual

• the true value of y – the estimated value of y

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Video 4: MR11.3 & 11.4

Properties of least squares

Hypothesis testing on the slope

•
• Make B1,0 = 0
Hypothesis testing on the intercept:

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Standard error of slope

Standard error of intercept

Video 5: MR11.4-11.7
Confidence intervals: distribution of regression coefficients

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Mean for the model

Mean when x = ….

•
Regression prediction intervals

•
Coefficient of determinations

Downloaded by ciko miko ([email protected])

 lOMoARcPSD|6274171

Downloaded by ciko miko ([email protected])