Scribd Logo
Upload

Welcome to Scribd!

  • 5 views

    iQRM Warm Up Week 5 February 17 Corrected

    Uploaded by

    mdorneanu

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    iQRM Warm Up Week 5 February 17 Corrected

    Uploaded by

    mdorneanu
    5 views39 pages

    Original Title

    iQRM Warm Up Week 5 February 17 corrected

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    5 views39 pages

    iQRM Warm Up Week 5 February 17 Corrected

    Uploaded by

    mdorneanu

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 39

    Spring 2024 Saturdays 8:30 - 9:00 am

    Instructor: Dr. Oleksandra (Sasha) Lyulko

    1
    iQRM Warm-Up: Last Week Review

    • Probability
    o Conditional probability of A given B is the probability of event A given that
    event B occurred: P(A)=P(A&B)/P(B)
    o Law of Total Probability: 𝑃 𝐴 = σ 𝑃 𝐴 𝐵 𝑃(𝐵)
    o Independent events A and B: P(A)=P(A|B)
    o Tree diagrams are useful to visualize conditional probability
    • Random Variables and Probability Distributions
    o Discreet RV: PMF
    o Continuous RV: PDF
    o Cumulative Distribution Function CMF: 𝐹𝑋 𝑥 = 𝑃(𝑋 ≤ 𝑥)
    o Distributions: Normal, Poisson, Lognormal, Binomial correspond to RV of
    different nature
    2
    iQRM Warm-Up: Today’s Topics

    • Measures of central tendency and variation


    o Mean, median, mode
    o Expected value
    o Standard deviation and variance
    • Normal Distribution

    3
    Measures of Central Tendency – Arithmetic Mean

    • Arithmetic mean – the average of all values of the random


    variable:
    𝑛
    1 𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛
    𝜇 = ෍ 𝑥𝑖 =
    𝑛 𝑛
    𝑖=1

    • Only valid when all values are equally likely!


    1+2+3+4+5+6
    • Fair die: 𝜇 = = 3.5
    6

    4
    What if the values are not equally likely?

    • Example: February is usually the


    snowiest month in New York City with
    snowfalls occurring on 3-4 days. Two
    of these days will record 1 inch (2.5
    cm) or more and 1 day will record 3
    inches (7.6 cm) or more.

    3 + 1 + 1 +0 + 0 + ⋯+ 0
    𝜇= ≈ 0.178
    28

    5
    Measures of Central Tendency – Expected Value

    3 + 1 + 1 +0 + 0 + ⋯+ 0 1 2 25
    𝜇= = ∙3+ ∙1+ ∙ 0 ≈ 0.178
    28 28 28 28

    • Weighted average!

    • 𝜇 = 𝑃(3) ∙ 3 + 𝑃(1) ∙ 1 + 𝑃(0) ∙ 0 ≈ 0.178

    6
    Measures of Central Tendency – Expected Value

    • Expected value, or expectation of a random variable with finite


    number of values is the weighted average of all possible
    values:
    𝐸 𝑋 = ෍ 𝑋 ∙ 𝑝(𝑋)

    • Expected value of a continuous random variable:

    𝐸 𝑋 = න 𝑥𝑓𝑋 𝑥 𝑑𝑥

    7
    Measures of Central Tendency – Median and Mode

    • Median - the value separating the upper half from the lower half
    of a data sample, a population, or a probability distribution.

    • Examples:
    • 5, 6, 8, 10, 12, 15, 20 – in case of odd number of data points, the
    median is the middle number
    • 5, 6, 8, 10, 12, 15, 20, 24 – in case of even number of data
    points the median is the average between the two middle
    numbers: 11

    8
    Measures of Central Tendency – Median and Mode

    • Mode - the most frequent number in a data set.

    • Examples:
    • 5, 6, 6, 10, 12, 15, 20 – mode is 6
    • Data sets can be multi-modal: 5, 6, 6, 10, 10, 15, 20 – modes
    are 6 and 10

    9
    Measures of Variation

    • Measures of variation are characteristics of distribution that


    show how far apart the data points are from each other or how
    spread the distribution is.

    • Range – the difference between the highest and lowest values


    • Interquartile range – the range of the middle half of the distribution
    • Variance – average of the squared distances from the mean
    • Standard deviation – square root of the variance

    10
    Measures of Variation – Variance

    • Variance – the average of the squared distances from the mean,


    i.e. expected value of the squared deviation from the mean:
    Var 𝑋 = 𝐸 𝑋 − 𝜇 2

    • Discrete random variable:


    • 𝑉𝑎𝑟 𝑋 = σ 𝑋 − 𝜇 2 ∙ 𝑝(𝑋)
    σ 𝑋−𝜇 2
    • 𝑉𝑎𝑟 𝑋 =
    𝑁

    • Continuous random variable: 𝑉𝑎𝑟 𝑋 = ‫ 𝑋(׬‬− 𝜇)2 𝑓𝑋 𝑥 𝑑𝑥

    11
    Measures of Variation - Variance

    • Question 1:
    • Why does the distribution
    with smaller variance
    have higher peak than the
    one with larger variance?
    • Answer:
    • Because the entire area
    (integral) should be equal
    to 1

    12
    Measures of Variation - Variance

    • Question 2:
    • Why can’t we simply find
    the average deviation?

    • Answer:
    • Because the average
    deviation is 0!

    13
    Properties of Expected Value and Variance

    • 𝐸 𝑎𝑋 + 𝑏𝑌 = 𝑎𝐸 𝑋 + 𝑏𝐸(𝑌)
    • Example: expected value of the sum of two dice

    • Var 𝑎𝑋 = 𝑎2 𝑉𝑎𝑟(𝑋)
    (if all values of the distribution are multiplied by 𝑎, squared devitions
    will be multiplied by 𝑎2 )

    • Var 𝑋 + 𝑌 = Var 𝑋 + Var(𝑌) – only when X and Y are independent


    • Variance of the mean of n independent, identically distributed variables:
    σ 𝑋𝑖 1 1 1
    • Var 𝑋ത = Var = Var σ 𝑋𝑖 = ∙ 𝑛 ∙ Var 𝑋 = Var 𝑋
    𝑛 𝑛2 𝑛2 𝑛
    14
    Population vs Sample

    • Population – the entire group of people, items or elements that


    we want to draw conclusions about
    • Sample – a subset of the population that we include in our
    study and collect data from

    15
    Population vs Sample

    16
    Population vs Sample

    17
    Measures of Variation – Standard Deviation

    • Standard deviation – square root of the variance:


    σ 𝑋−𝜇 2
    𝜎𝑋 = 𝑉𝑎𝑟(𝑋) =
    𝑁
    • Standard deviation of the sample when 𝝁 is not known:
    σ 𝑋 − 𝑋ത 2
    𝑠=
    𝑁−1
    𝜎(𝑋)
    • ത
    Standard deviation of the sample mean: 𝜎 𝑋 =
    𝑛

    18
    Parameters vs Statistics

    • Parameter is a number that describes population


    • Statistic is a number that describes sample

    Population parameters Sample statistics

    Population mean – 𝜇 Sample mean – 𝑋ത


    Population standard Sample standard
    deviation – 𝜎 deviation – 𝑠
    Population variance – 𝜎 2 Sample variance – 𝑠 2
    19
    Applications of Standard Deviation – Rule of 16

    • Volatility ~ standard deviation of the return.

    • Rule of 16: Given daily volatility, the annualized volatility can be


    calculated by multiplying by 16.

    • If annualized volatility is 80%, what is the daily volatility?

    20
    Other Measures – Skewness and Kurtosis

    • Skewness measures how symmetrical the distribution is


    around the mean.
    𝑬 𝑿−𝝁 𝟑
    𝒔=
    𝝈𝟑
    𝟏 𝒏 𝒙𝒊 −𝝁 𝟑
    or, for N data points observations: 𝒔 = σ𝒊=𝟏
    𝒏 𝝈

    21
    Other Measures – Skewness and Kurtosis

    • Positive skew:
    mean > median
    • Negative skew:
    mean < median

    22
    Other Measures – Skewness and Kurtosis

    • Kurtosis shows how widely spread the distribution is (similar to


    the variance):
    𝑬 𝑿−𝝁 𝟒
    𝑲=
    𝝈𝟒
    𝟏 𝒏 𝒙𝒊 −𝝁 𝟒
    or, for N data points observations: 𝑲 = σ𝒊=𝟏
    𝒏 𝝈

    23
    Excel Functions

    • Mean: =AVERAGE(select an array)


    • Median: =MEDIAN(array)
    • Mode: =MODE.SNGL(array), =MODE.MULT(array) (use with
    Ctrl+Shift+Enter)
    • Standard deviation: =STDEV.P(array), =STDEV.S(array)
    • Variance: =VAR.P(array), =VAR.S(array)

    24
    Probability Distributions – Normal Distribution

    • Most widely occurring in The population distribution of IQ scores drawn from


    a sample of 10,000 observations
    nature
    • Examples:
    • People’s heights
    • Measurement errors
    • SAT scores
    • IQ scores
    • Weight at birth

    From: “Learning Statistics with R” by Danielle Navarro 25


    History of Normal Distribution – Famous Contributors

    • 1783: Abraham de Moivre in his study “The Doctrine of Chances” approximated


    the coefficients of binomial expansion (𝑎 + 𝑏)𝑛 . The approximation curve was
    bell-shaped.
    • 1801-1823: Carl Friedrich Gauss discovered normal distribution. He
    introduced the two-parameter exponential function in 1809. Within
    astronomical studies of the dwarf planet Ceres, he proved the method of
    least squares under the assumption of normally distributed
    measurement errors and used the method to calculate the orbit of Ceres.
    2
    • 1782: Pierre-Simon Laplace calculated the integral ‫ 𝑒 ׬‬−𝑡 𝑑𝑡 = 𝜋 that
    normalizes Normal Distribution. In 1810 he proved Central Limit Theorem.
    • 1859-1866: James Clerk Maxwell developed kinetic theory of gases,
    where he derived the distribution of velocities of particles and
    demonstrated that Normal Distribution occurs in natural phenomena.

    26
    Normal Distribution - Definition

    • Normal Distribution (Gaussian Distribution) – continuous


    distribution for a real-valued random variable with probability
    density function (PDF)
    1 1 𝑥−𝜇 2
    −
    𝑓 𝑥 = 𝑒 2 𝜎
    𝜎 2𝜋

    27
    Normal Distribution – Probability Density Function

    1 1 𝑥−𝜇 2
    −
    𝑓 𝑥 = 𝑒 2 𝜎
    𝜎 2𝜋

    28
    Normal Distribution

    • Notation: 𝑋~𝑁(𝜇, 𝜎 2 ) means “random variable 𝑋 is drawn from


    a normal distribution with mean 𝜇 and standard deviation 𝜎”
    • How is it derived? The distribution of sample means
    approaches normal distribution as 𝑛 → ∞ (Central Limit
    Theorem – textbook pp. 73-76)
    • Example: distribution of log returns (𝑟 = ln(1 + 𝑅))
    • Properties: if 𝑋 and 𝑌 are normal random variables with means 𝜇𝑋
    and 𝜇𝑌 and standard deviations 𝜎𝑋 and 𝜎𝑌 , and 𝑍 = 𝑎𝑋 + 𝑏𝑌, then
    𝑍 is also a normal random variable with mean 𝜇𝑍 = 𝑎𝜇𝑋 + 𝑎𝜇𝑌 and
    variance 𝜎𝑍2 = 𝑎2 𝜎𝑋2 + 𝑏 2 𝜎𝑌2
    29
    Standard Normal Distribution

    • Standard Normal Distribution - normal distribution with the


    mean of 0 and standard deviation of 1
    𝑋−𝜇
    • If 𝑋~𝑁(𝜇, 𝜎 2 ), then 𝑍 = ~𝑁 0,1 (standard normal)
    𝜎

    • To get normal variable from standard normal: 𝑋 = σ𝑍 + 𝜇

    30
    What percent of the data lies within 1, 2 or 3
    standard deviations of the mean?

    • Integrate the PDF of Standard Normal Distribution from 0 to 1 to


    find what percent of the data lies between the mean (0) and
    mean + one standard deviation (1):
    1 −𝑥 2
    𝑓 𝑥 = 𝑒 2
    2𝜋

    31
    What percent of the data lies within 1, 2 or 3
    standard deviations of the mean?

    32
    Normally Distributed Data

    33
    Applying Normal Distribution

    • Example:
    • Although there is no law for IQ cutoff to become a police officer, in
    some states the police departments are allowed not to hire people
    who are too intelligent. Suppose in the Happytown Police
    Department the cutoff IQ score is 2 standard deviations above the
    population mean. Among 1000 people chosen at random,
    approximately how many will NOT get hired because they are too
    smart?

    • Answer: 2.2% of 1000 people are above 2 standard deviations ~ 22


    persons
    34
    Normal Distribution – Excel Functions

    • Probability of a value under a


    specific normal distribution:
    • Probability of a number given
    standard normal distribution:
    • Inverse of the normal
    cumulative distribution (given
    mean and stdev, what number
    is higher than the specified % of
    the data?):
    • Inverse of standard normal
    cumulative distribution:
    35
    Summary

    • We use mean (expected value), median and mode as measures


    of central tendency
    • We use variance and standard deviation as measures of variation
    • Normal distribution is bell-shaped and can be derived by finding
    the means of many samples
    • There is a specific percentage of data within 1, 2 and 3 standard
    deviation of the mean

    36
    Next Time

    • Hypothesis testing

    37
    Thank you!

    38
    Thank you! Any questions?

    You might also like