Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Chapter 8 Conditional and Binomial Probabilities D. S. Malik Creighton University, Omaha, NE D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 1 / 28 Conditional Probabilities Suppose that two distinct fair dice are rolled. We want to …nd the probability of getting a sum of 7, when it is given that the digit in the …rst die is greater than that of the second. Let S be the sample space of this experiment. Then n (S ) = 36. Because the dice are fair, each of these outcomes is equally likely. Let E be the event that the sum of the digits of the two dice is 7, and F be the event that the digit in the …rst die is greater than the second. Then E = f(6, 1), (5, 2), (4, 3), (3, 4), (2, 5), (1, 6)g F = f(6, 1), (6, 2), (6, 3), (6, 4), (6, 5), (5, 1), (5, 2), (5, 3), (5, 4), (4, 1), (4, 2), (4, 3), (3, 1), (3, 2), (2, 1)g. Let G be the event that the sum of the digits in the two dice is 7 but the digit in the …rst die is greater than the second. Then G = f(6, 1), (5, 2), (4, 3)g = E \ F . D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 2 / 28 It is given that the event F has occurred. Therefore, we reduce the original sample space of 36 outcomes to the sample space F . In this sample space, we consider the event G and the probability of G in F is n (G ) 3 1 = = . n (F ) 15 5 This is a conditional probability because it is given that the digit in the …rst die is greater than that of the second. We denote it by the symbol Pr[E j F ]. Now F and G are events in S. Thus, Pr[F ] = n (F ) 15 5 n (G ) 3 1 = = , Pr[G ] = = = , and n (S ) 36 12 n (S ) 36 12 Pr[G ] Pr[E \ F ] = = Pr[F ] Pr[F ] 1 12 5 12 1 = . 5 Hence, Pr[E j F ] = Pr[E \ F ] . Pr[F ] D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 3 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 4 / 28 De…nition Let E and F be two events in a sample space. Then the conditional probability of E , when it is given that the event F has occurred, i.e., Pr[F ] 6= 0, written Pr[E j F ], is given by Pr[E j F ] = Pr[E \ F ] . Pr[F ] Example Let E and F be events of an experiment. Suppose that Pr[E ] = 0.6, Pr[F ] = 0.7 and Pr[E \ F ] = 0.3. Then Pr[E j F ] = Pr[E \ F ] 0.3 3 = = . Pr[F ] 0.7 7 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 5 / 28 Example Let E and F be events of an experiment. Suppose that Pr[E ] = 0.65, Pr[F ] = 0.75 and Pr[E [ F ] = 0.9. We want to determine Pr[E j F ]. We are given Pr[F ], but not Pr[E \ F ]. So …rst we …nd Pr[E \ F ]. Now ) ) ) ) ) Pr[E [ F ] = Pr[E ] + Pr[F ] Pr[E \ F ] 0.9 = 0.65 + 0.75 Pr[E \ F ] 0.9 = 1.40 Pr[E \ F ] 0.9 1.40 = Pr[E \ F ] 0.50 = Pr[E \ F ] Pr[E \ F ] = 0.5. Hence, Pr[E j F ] = Pr[E \ F ] 0.5 50 2 = = = . Pr[F ] 0.75 75 3 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 6 / 28 Let E and F be two events of an experiment. Suppose that Pr[F ] 6= 0 and we want to …nd Pr[E j F ]. Now Pr[E j F ] is the probability of E given that the event F has occurred. We can think of this as a two-step experiment in which the …rst step consists of the occurrence of the event F and the second step consists of the occurrence of the event E . This means that we can draw a tree diagram showing the conditional probabilities as follows: F Pr[ ] Pr[ F’ ] | F] Pr[E E Pr[E’ E’ F | Pr[E F’ Pr[E’ | F] F’] E | F’] E’ Pr[F E] = Pr[FE] Pr[F E’] = Pr[FE’] Pr[F’ E] = Pr[F’E] Pr[F’ E’] = Pr[F’E’] D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities F E = FE F E’ = FE’ F’ E = F’E F’ E’ = F’E’ 7 / 28 Remark Let E and F be events of a sample space such that Pr[F ] 6= 0. From the previous …gure, it follows that Pr[E j F ] + Pr[E 0 j F ] = 1. This implies that Pr[E 0 j F ] = 1 Pr[E j F ]. D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 8 / 28 Suppose that two cards without replacement are drawn at random from a deck of 52 cards. Let S be the sample space of this event. We want to …nd Pr[2nd card is a kingj 1st card is a king]. Let F be the event that the …rst card is a king and E be the event that the second card is a king. Then F \ E is the event that the …rst card is a king and the second card is a king. Because two cards are drawn without replacement, n (S ) = 52 51 = 2652. Now F is the event that the …rst card is king. So the second card can be any one of the remaining 51 cards. Because the number of kings is 4, n (F ) = 4 51 = 204.Hence, Pr[F ] = 4 51 4 1 = = . 52 51 52 13 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 9 / 28 Next, F \ E = E \ F is the event that the …rst card is a king and the second card is a king. Because there are 4 kings and cards are drawn without replacement, it follows that n (E \ F ) = n (F \ E ) = 4 3 = 12. Thus, Pr[E \ F ] = 1 4 3 = . 52 51 221 Next we …nd Pr[E j F ]. This is a conditional probability. The sample space reduces to F . In the set F , the number of cards such that the second card is a king is 4 3 = 12. Thus, Pr[E j F ] = 4 3 4 3 3 1 = = = . n (F ) 4 51 51 17 You can also draw a tree diagram of this experiment as shown in the next slide: D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 10 / 28 |F P r[E F] Pr[ Pr[ F ’] =4 =4 / 52 8/5 2 51 ] = 3/ E F Pr[E’ F’ | P r[ E Pr[E’ | F] = 48/51 F’ ] = | F’] 4/51 = 47/ 51 E’ (4/52)(3/51) = 1/221 (4/52)(48/51) = 16/221 F (48/52)(4/51) =16/221 E E’ F (48/52)(47/51) = 188/221 E = FE E’ = FE’ F’ E = F’E F’ E’ = F’E’ Remark When drawing the tree diagram of conditional probabilities you must correctly identify the steps of the experiment and the outcomes. For a good number of problems, these steps can be easily identi…ed. D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 11 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 12 / 28 Example An urn contains 5 red, 6 brown, and 3 white marbles. Two marbles without replacements are drawn. What is the probability that the second marble is brown, given that the …rst marble is white? This can be considered a two-step experiment and we can draw the tree diagram as shown in the following diagram: D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 13 / 28 First Marble Second Marble (5/14)(4/13) = 10/91 R R = RR R 4/13 6/13 R 3/13 5/14 W R 5/13 6/14 5/13 B 3/13 B 5/13 6/13 (5/14)(3/13) = 15/182 R B = RB R W = RW B R = BR (6/14)(5/13) = 15/91 B B = BB (6/14)(5/13) = 15/91 (6/14)(3/13) = 9/91 W = BW R (3/14)(5/13) = 15/182 W R = WR B 2/13 Pr[2nd marble brown | 1st marble is white] (5/14)(6/13) = 15/91 B W 3/14 W B W (3/14)(6/13) = 9/91 (3/14)(2/13) = 3/91 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities W B = WB W W = WW 14 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 15 / 28 Theorem Let E and F be two events of an experiment. (i) Suppose that Pr[F ] 6= 0. Then Pr[E \ F ] = Pr[F ] Pr[E j F ]. (ii) Suppose that Pr[E ] 6= 0. Then Pr[E \ F ] = Pr[E ] Pr[F j E ]. De…nition Let E and F be two events of an experiment. Then E and F are called independent if Pr[E \ F ] = Pr[E ] Pr[F ]. Theorem Let E and F be two independent events of an experiment. (i) If Pr[F ] 6= 0, then Pr[E j F ] = Pr[E ], D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 16 / 28 Example Let E and F be events. (i) Suppose that Pr[E ] = 0.45, Pr[F ] = 0.48, and Pr[E \ F ] = 0.216. Now Pr[E ] Pr[F ] = 0.45 0.48 = 0.216 = Pr[E \ F ]. Thus, E and F are independent. (ii) Suppose that Pr[A] = 0.55, Pr[B ] = 0.62, and Pr[A \ B ] = 0.35. Now Pr[A] Pr[B ] = 0.55 0.62 = 0.341 6= 0.35 = Pr[A \ B ]. Thus, A and B are not independent. D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 17 / 28 Example To exercise regularly, a survey of 100 students shows that 40% play basketball, 70% jog, and 82% play either basketball or jog. Let B be the event that a randomly selected student plays basketball and J be the event that a randomly select student jogs. Note that Pr[B ] = 0.40, Pr[J ] = 0.70, and Pr[B [ J ] = 0.82. Now ) ) ) ) ) Pr[B [ J ] = Pr[B ] + Pr[J ] Pr[B \ J ] 0.82 = 0.40 + 0.70 Pr[B \ J ] 0.82 = 1.10 Pr[B \ J ] 0.82 1.10 = Pr[B \ J ] 0.28 = Pr[B \ J ] Pr[B \ J ] = 0.28. Thus, Pr[B ] Pr[J ] = 0.40 0.70 = 0.28 = Pr[B \ J ]. Hence, the events B and J are independent. D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 18 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 19 / 28 Exercise: Let E and F be events of a sample space such that Pr[E ] = 0.56, Pr[F ] = 0.45, and Pr[E \ F ] = 0.20. Find the following probabilities. (a) Pr[E j F ] (b) Pr[E 0 \ F ] Pr[F j E ] (e) Pr[E \ F 0 ] Solution: a. Pr[E j F ] = (c) Pr[E 0 j F ] (f) Pr[F 0 j E ] (d) 0.20 4 Pr[E \ F ] = = . Pr[F ] 0.45 9 b. Pr[E 0 \ F ] = Pr[F \ E 0 ] = Pr[F ] = 0.45 0.20 = 0.25. D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities Pr[E \ F ] 20 / 28 Pr[E ] = 0.56, Pr[F ] = 0.45, and Pr[E \ F ] = 0.20. Pr [E 0 \F ] 0.25 Solution: c. Pr[E 0 j F ] = Pr [F ] = 0.45 = 59 .You can also …nd Pr[E 0 j F ] by using the fact that Pr[E 0 j F ] = 1 Pr[E j F ] and part (a). That is, Pr[E 0 j F ] = 1 d. Pr[F j E ] = Pr[E j F ] = 1 5 4 = . 9 9 Pr[F \ E ] 0.20 5 = = . Pr[E ] 0.56 14 e. Pr[E \ F 0 ] = Pr[E ] f. Pr[F 0 j E ] = Pr[E \ F ] = 0.56 0.20 = 0.36. Pr[F 0 \ E ] 0.36 9 = = . Pr[E ] 0.56 14 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 21 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 22 / 28 Exercise: Solution: D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 23 / 28 Exercise: On the weekend at a shopping complex food plaza, 65% of the customers buy lunch. Of those who buy lunch, 20% also buy ice cream for a dessert. Of those who do not buy lunch, 55% buy ice cream for a dessert. a. What is the probability that a randomly selected customer does not buy ice cream, given that the customer does not buy lunch. b. What is the probability that a randomly selected customer buys ice cream. Solution: Let L be the set of customers who buy lunch and C be the set of customers who buy ice cream. Then from the given information Pr[L] = 0.65, Pr[C j L] = 0.20, and Pr[C j L0 ] = 0.55. Also, Pr[L0 ] = 1 Pr[L] = 1 0.65 = 0.35. a. We want to …nd Pr[C 0 j L0 ]. To …nd Pr[C 0 j L0 ], we use the fact that Pr[C 0 j L0 ] = 1 Pr[C j L0 ]. Thus, Pr[C 0 j L0 ] = 1 Pr[C j L0 ] = 1 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 0.55 = 0.45. 24 / 28 Solution: b. We want to …nd Pr[C ]. Note that C is the set of customers who buy ice cream. This is the same as the set of customers who buy lunch and ice cream, or the customers who do not buy lunch, but buy ice cream. That is C = (C \ L ) [ (C \ L0 ). So next we …nd Pr[C \ L]. Now Pr[C j L] = Pr[C \ L] . Pr[L] This implies that Pr[C \ L] = Pr[C j L] Pr[L] = (0.20) 0.(65) = 0.13. Similarly, Pr[C \ L0 ] = Pr[C j L0 ] Pr[L0 ] = (0.55) 0.(35) = 0.1925. Hence, Pr[C ] = Pr[C \ L] + Pr[C \ L0 ] = 0.13 + 0.1925 = 0.3225. D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 25 / 28 The previous exercise can also be answerd using a tree diagram as follows: 0.65 0.35 0.20 | L ]= Pr[C L Pr[C’ | L ]= 0.80 C C’ (0.65)(0.20) = 0.13 (0.65)(0.80) = 0.52 5 = 0.5 C (0.35)(0.55) = 0.1925 | L’ ] C [ r P L’ Pr[C’ (0.35)(0.45) = 0.1575 | L’ ] = 0.4 C’ 5 L C = LC L C ’= LC’ L’ C = L’C L’ C’ = L’C’ a. Pr[C 0 j L0 ] = 0.45. b. Pr[C ] = Pr[fLC , L0 C g] = Pr[LC ] + Pr[L0 C ] = 0.13 + 0.1925 = 0.3225 D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 26 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 27 / 28 Empty Slide for Notes D. S. Malik Creighton University, Omaha, NEChapter () 8 Conditional and Binomial Probabilities 28 / 28