The Negative Binomial and Uniform Distributions - PowerPoint PPT Presentation

1 / 21
About This Presentation
Title:

The Negative Binomial and Uniform Distributions

Description:

Each trial is a success' or failure' Probability of success is constant. The trials continue until a total of r successes have been observed ... – PowerPoint PPT presentation

Number of Views:718
Avg rating:3.0/5.0
Slides: 22
Provided by: alanab2
Category:

less

Transcript and Presenter's Notes

Title: The Negative Binomial and Uniform Distributions


1
The Negative Binomial and Uniform Distributions
  • MATH 224

2
Negative Binomial Distribution
  • Assumptions
  • Sequence of independent trials
  • Each trial is a success or failure
  • Probability of success is constant
  • The trials continue until a total of r successes
    have been observed
  • Measured the number of failures that precede the
    r-th success
  • Success can be misleading term

3
Negative Binomial Distribution
  • Contrast with binomial (non-negative version)

4
Examples
5
Distribution of Negative Binomial
6
Example the 100-year flood
  • A hundred-year flood is one which
  • happens every one hundred years
  • happens, on average, every 100 years
  • happens with probability 1/100 in each year

7
Probability of occurrence
  • You are building a flood-control system for a
    city
  • System is to last 30 years
  • Do you need to plan to handle a 1-in-100 year
    flood?
  • Do you need to plan for 2?

8
In MATLAB
p 1/100 (success 1-in-100 flood) r
1 x 130 for design, x 1( gt 100)
for interest x 1200 Px nbinpdf(x, r,
p) bar(x, Px) Fx nbincdf(x, r, p) bar(x,
Fx)
9
Graphs
10
Interpretation pdf vs cdf
11
Interpretation
  • Probability of a single 1-in-100 year flood, in
    100 years, is ___________
  • In a 30-year design, what is the probability of
  • one such flood?
  • two such floods?
  • one or more such floods?

12
Continued
13
Continuous vs Discrete
  • Binomial and negative binomial are discrete
    distributions because events are discrete
  • of pumps, flips, heads are all integers
  • Other events are continuous
  • rainfall, height, bending, concentration, time
  • Some are almost continuous
  • marks, age in years, date of birth

14
Use of continuous distributions
  • Continuous distributions are often used in
    discrete cases if there are a large number of
    possible events
  • Consider probability distribution of heads when
    tossing N coins for
  • N 10
  • N 100
  • N 1000

15
Graphical Comparison
16
Probability Density Functions
  • Discrete distributions have a probability of each
    event
  • P(6 working pumps) 0.9
  • In continuous distributions
  • probability of any exact event value is zero

17
Probability Density Functions 2
  • Probability is assigned to intervals
  • Quick reference the uniform distribution
  • Pick number evenly from interval 0, 10
  • Probability of 7.6827592875287 is_____
  • Probability of picking a number between 0 and 2
    is______________

18
Graphically
  • We draw the continuous version of a histogram,
    f(x), which is uniform over the interval 0, 10

19
Properties of continuous PDF
  • Given this probability density f(x),
  • Probability of x5 exactly is
  • Total probability, over all x values, is
  • P(2 lt x lt 4) can be computed using

20
Cumulative Distributions
  • Values like P(2ltxlt4) can be computed using same
    cumulative strategy we used for discrete
    distributions
  • Define uniform CDF, F(x) as

21
CDF uses
  • For uniform distribution on 0, 10, probability
    that a value between 2 and 4 will come up is
    ___________
  • More interesting with more interesting
    distributions (non-uniform)
Write a Comment
User Comments (0)
About PowerShow.com