Lecture On Probability Distributions

Similar documents
ELEG 3143 Probability & Stochastic Process Ch. 2 Discrete Random Variables

Example 1. The sample space of an experiment where we flip a pair of coins is denoted by:

The random variable 1

Chapter (4) Discrete Probability Distributions Examples

Probabilistic Systems Analysis Spring 2018 Lecture 6. Random Variables: Probability Mass Function and Expectation

CS 361: Probability & Statistics

Lecture 10: Probability distributions TUESDAY, FEBRUARY 19, 2019

Preliminary Statistics Lecture 2: Probability Theory (Outline) prelimsoas.webs.com

1 INFO Sep 05

Basic Probability space, sample space concepts and order of a Stochastic Process

Discrete Random Variables

success and failure independent from one trial to the next?

Chapter 3. Chapter 3 sections

MAT X (Spring 2012) Random Variables - Part I

Probability Theory. Introduction to Probability Theory. Principles of Counting Examples. Principles of Counting. Probability spaces.

Probability Rules. MATH 130, Elements of Statistics I. J. Robert Buchanan. Fall Department of Mathematics

Discrete Random Variables

(Ch 3.4.1, 3.4.2, 4.1, 4.2, 4.3)

IAM 530 ELEMENTS OF PROBABILITY AND STATISTICS LECTURE 3-RANDOM VARIABLES

Probability Distributions

(Ch 3.4.1, 3.4.2, 4.1, 4.2, 4.3)

Probability Methods in Civil Engineering Prof. Dr. Rajib Maity Department of Civil Engineering Indian Institution of Technology, Kharagpur

M378K In-Class Assignment #1

Discrete Mathematics and Probability Theory Spring 2014 Anant Sahai Note 10

Chapter 3: Probability 3.1: Basic Concepts of Probability

Session 2: Probability distributionsand density functions p. 1

Probability and Probability Distributions. Dr. Mohammed Alahmed

Homework 4 Solution, due July 23

Mixture distributions in Exams MLC/3L and C/4

Lecture 6. Probability events. Definition 1. The sample space, S, of a. probability experiment is the collection of all

Human-Oriented Robotics. Probability Refresher. Kai Arras Social Robotics Lab, University of Freiburg Winter term 2014/2015

Lecture 1: Probability Fundamentals

Statistics for Economists Lectures 6 & 7. Asrat Temesgen Stockholm University

STAT2201. Analysis of Engineering & Scientific Data. Unit 3

STA Module 4 Probability Concepts. Rev.F08 1

Notes for Math 324, Part 19

Probability Dr. Manjula Gunarathna 1

Probability Basics. Part 3: Types of Probability. INFO-1301, Quantitative Reasoning 1 University of Colorado Boulder

Review of Probability. CS1538: Introduction to Simulations

CME 106: Review Probability theory

Joint Probability Distributions and Random Samples (Devore Chapter Five)

Random Variable. Discrete Random Variable. Continuous Random Variable. Discrete Random Variable. Discrete Probability Distribution

Bernoulli Trials, Binomial and Cumulative Distributions

Random variables, Expectation, Mean and Variance. Slides are adapted from STAT414 course at PennState

CS1512 Foundations of Computing Science 2. Lecture 4

TOPIC 12 PROBABILITY SCHEMATIC DIAGRAM

Expectation of geometric distribution

Lecture 13: Covariance. Lisa Yan July 25, 2018

Part 3: Parametric Models

Binomial and Poisson Probability Distributions

Analysis of Engineering and Scientific Data. Semester

Lecture 2 Binomial and Poisson Probability Distributions

Discrete Random Variables

ECE 340 Probabilistic Methods in Engineering M/W 3-4:15. Lecture 2: Random Experiments. Prof. Vince Calhoun

Chapter 14. From Randomness to Probability. Copyright 2012, 2008, 2005 Pearson Education, Inc.

To find the median, find the 40 th quartile and the 70 th quartile (which are easily found at y=1 and y=2, respectively). Then we interpolate:

Probability Methods in Civil Engineering Prof. Dr. Rajib Maity Department of Civil Engineering Indian Institute of Technology Kharagpur

STAT 430/510 Probability Lecture 7: Random Variable and Expectation

Topic 3: The Expectation of a Random Variable

Discrete Probability Distributions

MA 250 Probability and Statistics. Nazar Khan PUCIT Lecture 15

Great Theoretical Ideas in Computer Science

Expectations and Variance

Binomial Distribution. Collin Phillips

Guidelines for Solving Probability Problems

Basic Probability. Introduction

Notes 12 Autumn 2005

Notes for Math 324, Part 17

Introduction and Overview STAT 421, SP Course Instructor

6.042/18.062J Mathematics for Computer Science November 28, 2006 Tom Leighton and Ronitt Rubinfeld. Random Variables

Bandits, Experts, and Games

Statistics Part I Introduction. Joe Nahas University of Notre Dame

Lecture 10: Bayes' Theorem, Expected Value and Variance Lecturer: Lale Özkahya

Expectation of geometric distribution. Variance and Standard Deviation. Variance: Examples

Discrete Mathematics for CS Spring 2007 Luca Trevisan Lecture 20

Dr. Junchao Xia Center of Biophysics and Computational Biology. Fall /13/2016 1/33

Special distributions

X = X X n, + X 2

Lecture 1: Basics of Probability

1. If X has density. cx 3 e x ), 0 x < 0, otherwise. Find the value of c that makes f a probability density. f(x) =

Discrete Mathematics and Probability Theory Spring 2016 Rao and Walrand Note 14

Bernoulli and Binomial Distributions. Notes. Bernoulli Trials. Bernoulli/Binomial Random Variables Bernoulli and Binomial Distributions.

Lecture 14. Text: A Course in Probability by Weiss 5.6. STAT 225 Introduction to Probability Models February 23, Whitney Huang Purdue University

Probability Experiments, Trials, Outcomes, Sample Spaces Example 1 Example 2

Probability Density Functions and the Normal Distribution. Quantitative Understanding in Biology, 1.2

ECO220Y Continuous Probability Distributions: Uniform and Triangle Readings: Chapter 9, sections

3 Lecture 3 Notes: Measures of Variation. The Boxplot. Definition of Probability

SDS 321: Introduction to Probability and Statistics

Week 12-13: Discrete Probability

Expectations and moments

ACM 116: Lecture 2. Agenda. Independence. Bayes rule. Discrete random variables Bernoulli distribution Binomial distribution

STAT 479: Short Term Actuarial Models

TOPIC 12: RANDOM VARIABLES AND THEIR DISTRIBUTIONS

A Probability Primer. A random walk down a probabilistic path leading to some stochastic thoughts on chance events and uncertain outcomes.

What is Probability? Probability. Sample Spaces and Events. Simple Event

Chapter 8: An Introduction to Probability and Statistics

Lecture 4: Probability and Discrete Random Variables

Discrete Random Variable

Chapter 2: Discrete Distributions. 2.1 Random Variables of the Discrete Type

II. The Binomial Distribution

Transcription:

Lecture On Probability Distributions 1 Random Variables & Probability Distributions Earlier we defined a random variable as a way of associating each outcome in a sample space with a real number. In our dice rolling experiment, each of the 36 possible outcomes must to be associated with one of the numbers through 1. We call x the sum of two faces on the dice a random variable since it associates each outcome with a real number from to 1. < How many different sums, can possibly turn up on the roll of two dice? >,3,4,5,6,7,8,9,10,11,1 or eleven different numbers. Now, for awhile we will limit ourselves to random variables which can take on a finite number of values i.e., we will limit ourselves to discrete random variables... later we will include a discussion of continuous random variables when we talk about the normal distributions. 1.1. Probability Mass Functions: Now that we know that a random variable is a real-numbered representation of the outcomes of an experiment, the question is how we associate our measures of probability with different values of our random variables, Definition: A probability MASS function is a function which assigns a probability P( x) to each real number (x) with in the range of a discrete random variable x. Examples: For each roll of two dice (fair) the probability that they will sum to a given number is given by the table. {next slide} This ProbDistLec_rev 013.lwp Page 1 of 58

correspondence between a value of a random variable and a measure of probability for that value represents a probability function connecting the value of probability. Sum of Dice 3 4 5 6 7 8 9 10 11 1 Total Probability is: Probability 1/36 /36 3/36 4/36 5/36 6/36 5/36 4/36 3/36 /36 1/36 1 < What is P(5) =? 4/36. P(9)? = 4/36. What is the sum of the probabilities for all R.V.'s > (1) Now, whenever we can we like to express probability functions by means of compact formulas we'll see momentarily, how we can write the probability functions for some very important probability distributions. {next slide} ProbDistLec_rev 013.lwp Page of 58

percentage Probability Distribution for Sum of Two Dice 18 15 1 9 6 3 0 3 4 5 6 7 8 9 10 11 1 Sum of Dice 1.. CUMULATIVE PROBABILITY FUNCTION: A cumulative probability function describes the cumulative probability that a random variable takes on values less than a given value. Take the dice throwing example just presented < What is the range of values of the RV Xs? > -1 <What is the PR(Xs <5)? = P(x=) + P(x=3) + P(x=4) + P(x=5) = 1/36 + /36 + 3/36 + 4/36 = 10/36 F(5) i5 P(x i) 10 36. {next slide} ProbDistLec_rev 013.lwp Page 3 of 58

Cumulative Probability Distribution for Sum of Two Dice 100 percentage 80 60 40 0 0 3 4 5 6 7 8 9 10 11 1 Sum of Dice 1.3. Properties of a Probability Function To summarize, the probability function of a random variable has two basic properties: {next slide} Properties of a probability function: 0 P x~ x 1 The probability of any value of a random variable is in the range (0,1) inclusive P x ~ 10. all x The probabilities of all values of a random variable in the sample space sum to 1. 1.4. Features of probability functions < Now, when we roll two (fair) dice, what is the most likely sum of the two? > If you look at the probability distribution, you would guess 7 right? {next slide} ProbDistLec_rev 013.lwp Page 4 of 58

percentage Probability Distribution for Sum of Two Dice 18 15 1 9 6 3 0 3 4 5 6 7 8 9 10 11 1 Sum of Dice most likely (or "expected") value Well, your intuition is actually quite good, the outcome most expected is the one with the highest probability of occurring. This works for a symmetric distribution, but not necessarily for all distributions. Just as we defined the mean of an empirical distribution of data as roughly the center of the distribution, we can define the mean of a random variable roughly as the center of mass of the probability distribution of the random variable. Mathematically, we weight each possible outcome by its probability and sum those products. That sum is the expected value of the random variable: {next slide} ProbDistLec_rev 013.lwp Page 5 of 58

Features of Probability Functions: Measure of Central Location: E x~ x P x all x The expected value of a random variable is the weighted sum of all its possible values, with the weights being the probability of each value. So, let's see how the formula for expected value works in the dice tossing case: {next slide} Sum of Dice Probability Products 1/36 x 0.078 3 /36 3 x 0.056 4 3/36 4 x 0.083 5 4/36 5 x 0.111 6 5/36 6 x 0.139 7 6/36 7 x 0.167 8 5/36 8 x 0.139 9 4/36 9 x 0.111 10 3/36 10 x 0.083 11 /36 11 x 0.056 1 1/36 1 x 0.078 Total Probability is: 1 Expected Value = all x P x~ i xi 7. 0 Secondly we have the expression for the variance of a probability distribution: {next slide} ProbDistLec_rev 013.lwp Page 6 of 58

Measure of the Spread of the Distribution: ~ ~ V x E x x P x all x The variance of a random variable is the expected value of the weighted sum of the squared deviations from the mean of the probability distribution, with the weights being the probability of each value. 1.5. Proof of Tchebysheff's Theorem A while back we talked about Tchebysheff's Theorem and about how, for any distribution, it allowed us to develop conservative estimates of the probability of obtaining a value within k standard deviations of the mean. Now that we know a little more about probability distributions, we can actually prove that theorem. Here's the theorem: {next slide} Tschebysheff's Theorem: Given any probability distribution with mean the probability of obtaining a value within k standard deviations of the mean is at least 1-1/k. i.e., Pr ~ 1 x k 1. k Note that Tschebysheff's theorem also implies that the probability that we get a value more than k standard deviations away from the mean is 1/k. Proof: {next slide} We know from our definition of a probability density function that: ~ ~ V x E x x P x all x ProbDistLec_rev 013.lwp Page 7 of 58

. We can split our probability distribution into 3 parts corresponding to the areas (a) outside the k region and (b) inside the k region region 1 region region 3 {next slide} mean k k 3. Then we can decompose the variance as follows: x Px x P x x P x region 1 region region 3 4. Since all three of these terms are greater than zero, we can drop the second term, leaving us with: x P x x P x region 1 region 3 5. Now, since the absolute value of the deviation of x from the mean, is at least k for all terms, we can write: region 1 region 3 k P x k P x ProbDistLec_rev 013.lwp Page 8 of 58

or, dividing through by k 1 k P x region 1 region 3 P x probability that x k probability that x k Therefore, the probability that a random variable is greater than k standard deviations away from the mean of its probability distribution is less than or equal to 1/k. End of Proof of Tschebysheff's theorem. 1.6. More Rules of Expectations Here are some more rules of expectations associated with probability functions: {next slide} {next slide} More Rules of Expectations E k k where k is a constant V k 0 the variance of a constant is zero (duh!) E kx ~ k E x ~ V kx ~ k V x ~ E x y E x E y ~ ~ ~ ~ E x y E x E y if x & y are independent. ~ ~ ~ ~ V x y V x V y if x & y are independent. always positive ProbDistLec_rev 013.lwp Page 9 of 58

The Expectation of a linear transformation of a random variable. The linear transformation is: ~ y = a + b ~ x This implies that the expectation of the linear transformation of x is: E ~ y E a b E ~ x a be ~ x A particularly useful linear transformation: Take any random variable x with mean and standard deviation. Form the new random variable: ~ z ~ x or, rearranging, we have, ~ 1 z ~ x. Now, let's find the expectation and variance of our new random variable z: E ~ 1 z E x ~. 1 0. The expectation of our new random variable z is zero! ProbDistLec_rev 013.lwp Page 10 of 58

The variance of our transformed random variable is: V ~ 1 z V x~ 1 V x~ V 1 V x~ 1 1 0 (Expectations Rule) (Expectations Rule) So, our new random variable has a variance of 1. Notice that this standardized random variable, z, has a mean = 0 and a variance/standard deviation = 1 irrespective of the type of distribution from which the random variable comes. These properties of the standardized random variable, z, will prove to be extremely useful when we begin to work with the normal distribution. Indeed, much of what we will be doing for the rest of the course involves taking a random variable x and transforming it into a new random variable y that has known properties (such as = 1 and expected value = 0). 1.7. Summary Now, we re at the point in the course where we start collecting a lot of information directly from probability theory which will later be enormously useful when we approach statistical inference. But first, let s see where we ve been: 1. We started off with the notion of uncertainty of attempting to find out things about an unknown population on the basis of samples from that population. ProbDistLec_rev 013.lwp Page 11 of 58

. We defined a measure of our uncertainty and called it probability and we said that our probability measure ought to have certain properties. 3. Then we went on to deduce other features of our probability measure which follow directly from our Postulates. Now, it's time to start looking directly at specific probability distributions that have known and useful properties that we can use in statistical inference to reduce our uncertainty about characteristics of unknown populations. Binomial Distribution Repeated identical trials are called Bernoulli Trials if three conditions are satisfied: 1. each trial has two possible outcomes, denoted generically, s, for success and f, for failure.. the trials are independent; and 3. the probability of a success remains the same from trial to trial, called the success probability and denoted p. The binomial distribution is the probability distribution for the number of successes in a sequence of Bernoulli trials. Example: Mortality: Mortality tables enable actuaries to obtain the probability that a person at any particular age will live a specified number of years. Such probabilities, in turn, permit the determination of life-insurance premiums, retirement pensions, annuity payments, and related items of importance to insurance companies and others. According to tables provided by the U.S. National Center for Health Statistics in Vital Statistics of the United States, there is about an 80% chance ProbDistLec_rev 013.lwp Page 1 of 58