Introduction Probability Math 141 Introduction to Probability and Statistics Albyn Jones Mathematics Department Library 304 [email protected] www.people.reed.edu/∼jones/courses/141 September 3, 2014 Albyn Jones Math 141 Introduction Probability Motivation How likely is an eruption at Mount Rainier in the next 25 years? Albyn Jones Math 141 Introduction Probability Data! Post ice-age eruptions Mount Rainier vs a Poisson Point Process Poisson ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ●● ● ● ● ●● ● ●● ● ● ●● ●● ● ● ●●●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● Mount Rainier −8000 −6000 −4000 Year Albyn Jones Math 141 −2000 0 ● Introduction Probability Two Models Don’t worry about the details! Poisson Process: Events uniformly distributed in time. Roughly 50 events in the last 12000 years: one every 240 years. Albyn Jones Math 141 Introduction Probability Two Models Don’t worry about the details! Poisson Process: Events uniformly distributed in time. Roughly 50 events in the last 12000 years: one every 240 years. Prediction: roughly a 10% chance of an eruption in the next 25 years, regardless of the elapsed time since the last eruption. Albyn Jones Math 141 Introduction Probability Two Models Don’t worry about the details! Poisson Process: Events uniformly distributed in time. Roughly 50 events in the last 12000 years: one every 240 years. Prediction: roughly a 10% chance of an eruption in the next 25 years, regardless of the elapsed time since the last eruption. Hidden Markov Model: Two (unobservable) states with different rates. Given the last eruption was roughly 1050 years ago, we think we are in a low rate regime: roughly one eruption every 650 years. Albyn Jones Math 141 Introduction Probability Two Models Don’t worry about the details! Poisson Process: Events uniformly distributed in time. Roughly 50 events in the last 12000 years: one every 240 years. Prediction: roughly a 10% chance of an eruption in the next 25 years, regardless of the elapsed time since the last eruption. Hidden Markov Model: Two (unobservable) states with different rates. Given the last eruption was roughly 1050 years ago, we think we are in a low rate regime: roughly one eruption every 650 years. Prediction: roughly a 3.7% chance of an eruption in the next 25 years. Albyn Jones Math 141 Introduction Probability Questions Hint: statistical analysis! Which of those predictions is more reliable? In other words: which is the better model? Albyn Jones Math 141 Introduction Probability Questions Hint: statistical analysis! Which of those predictions is more reliable? In other words: which is the better model? How do we produce estimates for those models? Albyn Jones Math 141 Introduction Probability Questions Hint: statistical analysis! Which of those predictions is more reliable? In other words: which is the better model? How do we produce estimates for those models? How accurate or trustworthy are those estimates? Albyn Jones Math 141 Introduction Probability Questions Hint: statistical analysis! Which of those predictions is more reliable? In other words: which is the better model? How do we produce estimates for those models? How accurate or trustworthy are those estimates? How do we validate the models? Albyn Jones Math 141 Introduction Probability Statistics what is it all about? Formal inference: estimates, confidence intervals and hypothesis tests; quantification of uncertainty. Albyn Jones Math 141 Introduction Probability Statistics what is it all about? Formal inference: estimates, confidence intervals and hypothesis tests; quantification of uncertainty. Tools: probability theory, computational engines like R. Albyn Jones Math 141 Introduction Probability Statistics what is it all about? Formal inference: estimates, confidence intervals and hypothesis tests; quantification of uncertainty. Tools: probability theory, computational engines like R. Informal inference: judgements about statistical models — model choice, model validation Albyn Jones Math 141 Introduction Probability Statistics what is it all about? Formal inference: estimates, confidence intervals and hypothesis tests; quantification of uncertainty. Tools: probability theory, computational engines like R. Informal inference: judgements about statistical models — model choice, model validation Tools: graphical methods, computational engines like R. Albyn Jones Math 141 Introduction Probability Statistics what is it all about? Formal inference: estimates, confidence intervals and hypothesis tests; quantification of uncertainty. Tools: probability theory, computational engines like R. Informal inference: judgements about statistical models — model choice, model validation Tools: graphical methods, computational engines like R. Note the computational theme! Albyn Jones Math 141 Introduction Probability A Little Probability Theory The mathematics we need to quantify uncertainty A little History: gambling! dice! cards! A little Philosophy: epistemology and subjective probability, positivism and ‘objective’ probability. Albyn Jones Math 141 Introduction Probability Example Toss a fair coin 3 times. What is the probability that two of the three tosses yield heads? We need some terminology and notation: Sample Space: the set of possible outcomes. Albyn Jones Math 141 Introduction Probability Example Toss a fair coin 3 times. What is the probability that two of the three tosses yield heads? We need some terminology and notation: Sample Space: the set of possible outcomes. Event: a subset of the sample space. Albyn Jones Math 141 Introduction Probability Example Toss a fair coin 3 times. What is the probability that two of the three tosses yield heads? We need some terminology and notation: Sample Space: the set of possible outcomes. Event: a subset of the sample space. Probability: a function assigning real numbers to events. Albyn Jones Math 141 Introduction Probability Sample Space: Ω Toss a fair coin 3 times. What are the possible outcomes? {HHH} {HHT }, {HTH}, {THH} {HTT }, {THT }, {TTH} {TTT } These are the events in our sample space. Albyn Jones Math 141 Introduction Probability Notation A little set theory Let A and B be events (subsets of the sample space Ω). Term Notation Interpretation Union A∪B A or B occurs (or both!) Intersection A∩B A and B both occur Complement Ac , !A, (Ω \ A) A does not occur Disjoint Events A∩B =∅ A and B can not both occur Albyn Jones Math 141 Introduction Probability Notation: Examples three coin tosses again ‘two heads’: a union of three events {HHT } ∪ {HTH} ∪ {THH} Albyn Jones Math 141 Introduction Probability Notation: Examples three coin tosses again ‘two heads’: a union of three events {HHT } ∪ {HTH} ∪ {THH} ‘at least one head’: the complement of ‘no heads’ {TTT }c = {TTH} ∪ {THT } ∪ . . . ∪ {HHH} Albyn Jones Math 141 Introduction Probability Notation: Examples three coin tosses again ‘two heads’: a union of three events {HHT } ∪ {HTH} ∪ {THH} ‘at least one head’: the complement of ‘no heads’ {TTT }c = {TTH} ∪ {THT } ∪ . . . ∪ {HHH} an impossible event! {TTT } ∩ {HHH} Albyn Jones Math 141 Introduction Probability Probability three coin tosses again Let’s assign probabilities to our 8 events, giving each event in Ω the same probability (why?). P{HHH} = P{HHT } = . . . P{TTT } = Note the probabilities of the 8 events add up to 1. Albyn Jones Math 141 1 8 Introduction Probability More Probability! three coin tosses again Now, what is the probability of getting two heads in three tosses? 1 8 The probabilities of these 3 events add up to 3/8. Is that the correct value for the probability of getting two heads? P{HHT } = P{HTH} = P{THH} = We need some rules for computing probabilities! Albyn Jones Math 141 Introduction Probability Rules for Probability aka AXIOMS Let Ω be a sample space, and E1 , E2 , E3 , . . . be events. P : {Events} → R according to the following three rules: Albyn Jones Math 141 Introduction Probability Rules for Probability aka AXIOMS Let Ω be a sample space, and E1 , E2 , E3 , . . . be events. P : {Events} → R according to the following three rules: 1 For any event E: 0 ≤ P(E) ≤ 1 Albyn Jones Math 141 Introduction Probability Rules for Probability aka AXIOMS Let Ω be a sample space, and E1 , E2 , E3 , . . . be events. P : {Events} → R according to the following three rules: 1 For any event E: 0 ≤ P(E) ≤ 1 2 P(Ω) = 1 Albyn Jones Math 141 Introduction Probability Rules for Probability aka AXIOMS Let Ω be a sample space, and E1 , E2 , E3 , . . . be events. P : {Events} → R according to the following three rules: 1 For any event E: 0 ≤ P(E) ≤ 1 2 P(Ω) = 1 3 If E1 , E2 , E3 , . . . are disjoint events, then X P(E1 ∪ E2 ∪ . . .) = P(Ei ) = P(E1 ) + P(E2 ) + . . . Albyn Jones Math 141 Introduction Probability Example three coin tosses again Now, what is the probability of getting two heads in three tosses, given our assignment of equal probability to each: P({HHT }) = P({HTH}) = P({THH}) = Albyn Jones Math 141 1 8 Introduction Probability Example three coin tosses again Now, what is the probability of getting two heads in three tosses, given our assignment of equal probability to each: P({HHT }) = P({HTH}) = P({THH}) = Are these events disjoint? Albyn Jones Math 141 1 8 Introduction Probability Example three coin tosses again Now, what is the probability of getting two heads in three tosses, given our assignment of equal probability to each: P({HHT }) = P({HTH}) = P({THH}) = Are these events disjoint? yes! Albyn Jones Math 141 1 8 Introduction Probability Example three coin tosses again Now, what is the probability of getting two heads in three tosses, given our assignment of equal probability to each: P({HHT }) = P({HTH}) = P({THH}) = 1 8 Are these events disjoint? yes! Therefore, by axiom 3, P({HHT } ∪ {HTH} ∪ {THH}) = P({HHT }) + P({HTH}) + P({THH}) = Albyn Jones Math 141 3 8 Introduction Probability Complementary Events What do we know about the events E and E c ? What is E ∩ E c ? What is E ∪ E c ? Albyn Jones Math 141 Introduction Probability More on Complementary Events For any event E, E ∩ E c = ∅, so E and E c are disjoint. For any event E, E ∪ E c = Ω. Putting these facts together with our axioms: 1 = P(Ω) = P(E ∪ E c ) = P(E) + P(E c ) Thus P(E c ) = 1 − P(E) Albyn Jones Math 141 Introduction Probability Example: Complementary Events What is the probability of at least one head in three tosses? Albyn Jones Math 141 Introduction Probability Example: Complementary Events What is the probability of at least one head in three tosses? What is the complement of ‘at least one head’? Albyn Jones Math 141 Introduction Probability Example: Complementary Events What is the probability of at least one head in three tosses? What is the complement of ‘at least one head’? No heads! (All tails.) Albyn Jones Math 141 Introduction Probability Example: Complementary Events What is the probability of at least one head in three tosses? What is the complement of ‘at least one head’? No heads! (All tails.) Using the last result we have P(at least one head) = P({TTT }c ) = 1 − P({TTT }) = 1 − Albyn Jones Math 141 7 1 = 8 8 Introduction Probability A probability inequality If A ⊂ B, then P(A) ≤ P(B), proof by picture: A B Albyn Jones Math 141 Introduction Probability A General Addition Formula: Inclusion/Exclusion P(A ∪ B) = P(A) + P(B) − P(A ∩ B) B A Albyn Jones Math 141 Introduction Probability Summary 1 Definitions: Sample Space, Events, Disjoint Events 2 Axioms or Rules of Probability 3 P(E) = 1 − P(E c ) 4 Addition formula: P(A ∪ B) = P(A) + P(B) − P(A ∩ B) Albyn Jones Math 141 Introduction Probability Assignment! Read Chapter 2. Albyn Jones Math 141
© Copyright 2026 Paperzz