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 2 Concepts of Prob. Theory • Random Experiment: an experiment in which outcome varies in an unpredictable fashion when the experiment is repeated under the same condition Specified by: 1. An experimental procedure 2. One or more measurements or observations tch-prob 1 EXAMPLE 2.1 : Experiment E1 : Select a ball form an urn containing balls numbered 1 to 50. Note the number of the ball. Experiment E2 : Select a ball form an urn containing balls numbered 1 to 4. Suppose that balls 1 and 2 are black and that balls 3 and 4 are white. Note number and color of the ball you select. Experiment E3 : Toss a coin three times and note the sequence of heads/ tails. Experiment E4 : Toss a coin three times and note the number of heads. Experiment E6 : A block of information is transmitted repeatedly over a noisy channel until an error-free block arrives at the receiver. Experiment E7 : Pick a number at random between zero and one. Experiment E12 : Pick two numbers at random between zero and one. Experiment E13 : Pick a number X at random between zero and one, then pick a number Y at random between zero and X. tch-prob 2 Discussions • Compare E3 and E4: same procedure, different observations • A random experiment with multiple measurements or observations: E2, E3, E12, E13 • sequential experiment consists of multiple subexperiments: E3, E4, E6, E12, E13 • dependent subexperiments: E13 tch-prob 3 Sample Space • Outcome, sample point: – one and only one per experiment – mutually exclusive, cannot occur simultaneously • Sample space: set of all possible outcomes • Example 2.2: sample spaces • S1, S2 , S3, S6 , S7 , S12 , S13 number of outcomes: – Finite – countably infinite S6 – unaccountably infinite S7 tch-prob Discrete sample space Continuous Sample space 4 EXAMPLE 2.2 : The sample spaces corresponding to the experiments in Example 2.1 are given below using set notation : S1 1, 2,......,50 S2 1, b , 2, b , 3, w , 4, w S3 HHH , HHT , HTH , THH , TTH , THT , HTT , TTT S4 0,1, 2,3 S6 1, 2,3, S7 x : 0 x 1 0,1 S12 x, y : 0 x 1 and 0 y 1 S13 x, y : 0 y x 1 tch-prob 5 Event: a subset of S, a collection of outcomes that satisfy certain conditions Certain event: S null event: elementary event: event of a single outcome E ,E ,E 2 4 7 A2: The ball is white and even-numbered Example 2.3: A4: The number of heads equals the number of tails A7: The number selected is nonnegative tch-prob 6 Set Operations 1. 2. Union A B A Intersection A B mutually exclusive 3. B Complement imply A B equal A=B A B Ac S A - Commutative properties of set operations A BB A A BB - Associative properties of set operations A ( B C ) ( A B) C - Distributive properties of set operations A ( B C ) ( A B) ( A C ) - De Morgan’s Rules Venn diagrams: ( A B)c Ac Bc ( A B)c Ac A A ( B C ) ( A B) C Bc Fig. 2.2 tch-prob 7 Example 2.5: Configuration of a three-component system a. series all three C1 C2 C3 A1 I A2 I A3 b. parallel at least one of three A1 A2 C1 C2 A3 C3 c. two-out-of-three ( A1 A2 A3 ) ( A1c n A A A ..... An k 1 k 1 2 n A A A ..... An k 1 k 1 2 A A k k 1 k k 1 A2 A3 ) ... Ak : event that component k is functioning tch-prob 8 2.2 The Axioms of Probability A probability law for the experiment E is a rule that assigns to each event a number P(A), called the probability of A, that satisfies the following axioms: Axiom I: Axiom II: 0 P( A) nonnegative P(S)=1 Axiom III: If total=1 A B , then P A B P A P B Axiom III’: If Ai A j for all i j , then P A P A k k k 1 k 1 A set of consistency rules that any valid probability assignment must satisfy. tch-prob 9 Corollary 1. P Ac 1 P A pf: A Ac 1 P S P A Ac P A P Ac Corollary 2. P A 1 pf: from Cor.1, Corollary 3. P 0 pf: Let A=S, Corollary 4. P A 1 P Ac 1 Ac in Cor.1. A , A ,...., An, are pairwise mutually exclusive, 1 2 then n n P A P A for n 2 k 1 k k 1 k tch-prob 10 Corollary 5. P A B P A P B P A B pf: P A B P A Bc P B Ac P A B P A P A Bc P A B P B P B Ac P A B A B Corollary 6. P n n1 A P A P A A .... 1 P A .... An j j k 1 k 1 k j1 jk n Corollary 7. If A B , then P A P B pf. P B P A P Ac B P A tch-prob 11 Axioms + corollaries provide rules for computing the probability of certain events in terms of other events. However, we still need initial probability assignment 1. Discrete Sample Spaces find the prob.of elementary events; all distinct elementary events are mutually exclusive, Example 2.6., 2.7 2. Continuous Sample Spaces assign prob. to intervals of the real line or rectangular regions in the plane Example 2.11 tch-prob 12 Discrete Sample Spaces First, suppose that the sample space is finite, and is given by S a1 , a2 an P B P a1' , a2' , , am' P a1' P a2' (by corollary ?) P am' (2.9) If S is countably infinite, then Axiom Ⅲ’ implies that the probability of an event such as D b1' ,b2' is given by P D P b1' P b2' (2.10) tch-prob 13 If the sample space has n elements, S a1 ,an , a probability assignment of particular interest is the case of equally likely outcomes. The probability of the elementary events is Pa1 Pa2 Pan 1 n (2.11) The probability of any event that consists of k outcomes, say B a1' , ak' , is k PB Pa Pa n ' 1 ' k 2.12 Thus, if outcomes are equally likely, then the probability of an event is equal to the number of outcomes in the event divided by the total number of outcomes in the sample space. (remember the classical definition?) tch-prob 14 EXAMPLE 2.6 An urn contains 10 identical balls numbered 0,1,…, 9. A random experiment involves selecting a ball from the urn and noting the number of the ball. Find the probability of the following events : A = “number of ball selected is odd,” B = “number of ball selected is a multiple of 3,” C = “number of ball selected is less than 5,” and of A B and A B C . tch-prob 15 The sample space is S 0,1,,9 , so the sets of outcomes corresponding to the above events are A 1,3,5,7,9, B 3,6,9 , and C 0,1,2,3,4. If we assume that the outcomes are equally likely, then PA P1 P3 P5 P7 P9 PB P3 P6 P9 5 10 3 10 . PC P0 P1 P2 P3 P4 5 . 10 From Corollary 5. 5 3 2 6 PA B PA PB PA B . 10 10 10 10 tch-prob 16 where we have used the fact that A B 3,9, so PA B 2 10. From Corollary 6, PA B C PA PB PC PA B P A C P B C P A B C 5 3 5 2 2 1 1 10 10 10 10 10 10 10 9 10 tch-prob 17 EXAMPLE 2.7 Suppose that a coin is tossed three times. If we observe the sequence of heads and tails, then there are eight possible outcomes . If we S HHH , HHT , HTH , THH , TTH , THT , HTT , TTT assume that the outcomes of S3 are equiprobable, then the probability of each of the eight elementary events is 1/8. This probability assignment implies that the probability of obtaining two heads in three tosses is, by Corollary 3, P "2 heads in 3 tosses" PHHT , HTH , THH PHHT PHTH PTHH 3 8 *If we count the number of heads in three tosses, then S 0,1, 2,3 P "2 heads in 3 tosses" P2 tch-prob 1 4 18 Continuous Sample Spaces EXAMPLE 2.11 Consider Experiment E12, where we picked two number x and y at random between zero and one. The sample space is then the unit square shown in Fig. 2.8(a). If we suppose that all pairs of numbers in the unit square are equally likely to be selected, then it is reasonable to use a probability assignment in which the probability of any region R inside the unit square is equal to the area of R. Find the probability of the following events: A x 0.5, B y 0.5 , and C x y . Figures 2.8(b) through 2.8(c) show the regions corresponding to the events A, B, and C. Clearly each of these regions has area ½. 1 1 1 Thus PA PB PC 2 2 2 tch-prob 19 FIGURE 2.8 a Two-dimensional sample space and three events. y y x S 1 2 x x 0 0 1 (a) Sample space 1 (b) Event y 1 x 2 y y 1 2 x y 0 (c) Event 1 x x 0 1 (d) Event x y 1 y 2 tch-prob 20 2.3 Computing probabilities using counting methods In many experiments with finite sample space, the outcomes can be assumed to be equiprobable. Prob. 1. Counting the number of outcomes in an event Sampling with Replacement and with Ordering choose k objects from a set A that has n distinct objects, with replacement and note the order. The experiment produces an ordered k-tuple x x x 1, 2,..., k where xi A and i 1,2,..., k The number of distinct ordered k-tuples = nk Ex.2.12 5 balls, 1,2,3,4,5 select 2 52 25 Pr[2 balls with same number] = 5 25 tch-prob 21 2. Sampling without Replacement and with ordering choose k objects from a set A that has n distinct objects, without replacement. k n number of distinct ordered k-tuples = n (n-1)…(n-k+1) Ex.2.13 5 balls, select 2 5 4 20 Pr [1st number >2nd number ]= 10/20 tch-prob 22 Permutations of n Distinct Objects Consider the case when k=n number of permutations of n objects = n! For large n, stirling’s formula is useful n1 n! 2 n 2en Ex. 2.16 12 balls randomly placed into 12 cells, where more than 1 ball is allowed to occupy a cell. Pr [ all cells are occupied ]= 12! 12 5 5.37 10 ~ 12 tch-prob 23 3. Sampling without Replacement and without ordering n ~ ” combination of size k” = Ck each combination has k! possible orders C n k ! n n 1.... n k 1 k n n 1.... n k 1 n C n! k k! k ! n k ! k n “binomial coefficient”, read “n choose k” k n Ex. 2.18 The number of distinctive permutation of k white balls and n-k black balls n! k ! n k ! tch-prob 24 Ex. 2.19 A batch of 50 items contains 10 defective items. Select 10 items at random, Pr [ 5 out of 10 defective ]=? 10 40 5 5 50 10 Exercise prob. 47: Multinomial Ex.2.20 Toss a die 12 times. Pr [ each number appears twice ]= ? 12! 2!2!2!2!2!2! 612 {1,1,2,2,3,3, …….6,6} permutation tch-prob 25 4. Sampling with Replacement and without ordering n 1 k n 1 k k n 1 Pick k balls from an urn of n balls. (k may be greater than n) Suppose k=5, n=4 object 1 2 XX 3 4 X XX The result can be summarized as XX // X / XX 8 8! = k: x 3!5! 3 n-1: / k n 1! = k ! n 1 ! Ex. Place k balls into n cells. tch-prob 26 2.4 Conditional Probability We are interested in knowing whether A and B are related, i.e. , if B occurs, does it alter the likelihood of A occurs. Define the conditional probability as B P A| B P AI B P B - renormalize A for P[B]>0 P[ A I B ] tch-prob 27 Example 2.21. A ball is selected from an urn containing 2 black balls (1,2) and 2 white balls (3,4). A: Black ball selected ½ B: even-numbered ball ½ C: number > 2 ½ P A I B 0.25 P A B 0.5 P A 0.5 P B P A I C P AC 0 0 P A 0.5 P C tch-prob Not related Related 28 P AI B P B P A B is useful in finding prob. in sequential experiments P A P B A Ex.2.22 An urn contains 2 black, 3 white balls. Two balls are selected without replacement. Sequence of colors is noted. Sol. 1. b b two subexperiments B1 bw W1 2/5 3/5 wb B2 ww 1/4 P[B1 I B2 ] P[B1]P[B2 / B1] tch-prob 1/10 W2 3/4 3/10 B2 W2 2/4 3/10 2/4 3/10 Tree diagram 29 Let B1 , B2 , , Bn be mutually exclusive, whose union = S , S partition of S. A A I S A I (B1 U B2 UU Bn ) A I B1 U A I B2 U...U AI Bn B1 B2 A Bn P A P AI B1 P AI B2 ... P AI Bn “ Theorem on total probability” P A P B1 P A B1 P B2 P A B2 ... P A Bn P Bn Ex. 2.24 In prev. example (2.22) find P[W2]=Pr [second ball is white] P W2 P W2 B1 P B1 P W2 W1 P W1 3 2 1 3 3 4 5 2 5 5 tch-prob 30 Suppose event A occurs, what is the prob. of event Bj? Bayes’ Rule P B j AI B A B P B P P j j A j n P A P A Bk P Bk k 1 Before experiment: P[ Bj ]= a priori probability After experiment: A occurs P[ Bj/A ]= a posteriori probability tch-prob 31 Ai Ex.2.26. Binary symmetric channel 1-p 0 Suppose P A P A 1 0 1 2 1-e Bj 0 e e p 1 1-e 1 P B1 P B1 A0 P A0 P B1 A1 A1 e 1 1 e 1 1 2 2 2 B A P A P e/2 P A0 B1 1 0 0 1/ e, 2 P B1 P A1 B1 1 e tch-prob 32 2.5 Independence of Events If knowledge of the occurrence of an event B does not alter the probability of some other event A, then A is independent of B. P A I B A P A B P P B Two events A and B are independent if P AI B P A P B tch-prob 33 Ex. 2.28. A ball is selected from an urn of 4 balls {1,b},{2,b},{3,w},{4,w} A: black ball is selected B: even-numbered ball C: number >2 P[A]=P[B]=P[C]=0.5 P AI B 1 P A P B 4 P AI C 0 P A B P A P A C P A If two events have non zero prob., and are mutually exclusive, then they cannot be independent. 0 P AI B P A P B 0 tch-prob 34 Three events A, B, and C are independent if P A I B P A P B P A I C P A P C P B I C P B P C P A I B I C P A P B P C Ex. 2.30. P B I D 1 P B P D 4 P B I F 1 P B P F 4 P D I F 1 P D P F 4 P B I D I F 0 P B P D P F B F D F Common application of the independence concept is in making the assumption that the event of separate experiments are independent. Ex.2.31. A fair coin is tossed three times. P[HHH]=1/8, …. tch-prob 35 2.6. Sequential Experiments • Sequence of independent experiments If sub experiments are independent, and event Ak concerns outcomes of the kth subexperiment, then A1 , A2 , , An are independent. P A1 I A2 I I An P A1 P A2 ...P An • Bernoulli trial: Perform an experiment once, note whether an event A occurs. success, if A occurs prob= p failure, otherwise. prob= 1-p Pn(k)= prob. of k successes in n independent repetitions of a Bernoulli trial = ? tch-prob 36 Ex.2.34. toss a coin 3 times. P k 0 P[ H ]= p P k 2 3 p2 1 p 1 p 3 P k 1 3 p 1 p 2 P k 3 p3 Binomial probability law pn k n k pk 1 p nk k=0,1,…,n tch-prob 37 Binomial theorem n n k nk n a b a b k 0 k Let a=b=1 , Let n n n 2 k 0 k a p,b 1 p, n n k n nk 1 p 1 p p k n k 0 k k 0 Ex.2.37. A binary channel with BER 103 . 3 bits are transmitted. majority decoding is used. What is prob. of error? 3 3 2 P k 2 .001 .999 .0013 3106 2 3 tch-prob 38 Multinomial Probability Law P B j p j Bj’s: partition of S; mutually exclusive p1 p2 pM 1 n independent repetitions of experiment B1 B2 .... .... S BM k j number of times Bj occurs. P[(k1, k2 ,..., kM )] n! p1k1 p2k2 ... pM kM k1 !k2 !...kM ! Ex.2.38. dart 3 areas p 0.2,0.3,0.5 throw dart 9 times, P[3 on each area]=? P 3,3,3 9! 3 .3 3 .5 3 .2 3!3!3! tch-prob 39 Geometric Probability Law Repeat independent Bernoulli trials until the occurrence of the first success. p m 1 p ...1 p p 1 p m1 p, m1 m1,2,... times 1 m 1 p p m p 1 p 1 11 p m 1 m 1 P[ more than k trials are required before a success occurs ] p[m k ] p 1 p m11 p k mk 1 P[The initial k trials are failures] tch-prob 40 Sequence of Dependent Experiments Ex.2.41. A sequential experiment involves repeatedly drawing a ball from one of two urns, noting the number on the ball, and replacing the ball in its urn. Urn 0 : 1 ball #1, 2 balls #0 Urn 1 : 5 balls #1, 1 ball #0 Initially, flip a coin. Thereafter, the urn used in the subexperiment corresponds to the number selected in the previous subexperiment. 0 0 0 0 0 h 1 1 1 0 .... Urn 2 0 1 S Urn 1 1 0 t .... .... 1 1 S 1 S 2 S 3 Trellis diagram tch-prob 41 P S S P S S P S conditional prob. 0 1 1 0 0 PS S S PS S S PS S 0 1 2 2 0 1 0 1 PS S S PS S PS 2 0 1 1 0 0 PS S PS S PS Markov property 2 1 1 0 0 P S0 S1 ,... Sn P Sn Sn1 P Sn1 Sn2 ...P S1 S0 P S0 .... Ex.2.42 calculate with P 0011 P 1 1 P 1 0 P 0 0 P 0 1 2 1 P1 0 P 0 0 P 0 P[1] 3 3 2 1 5 P 1 1 P 0 1 6 6 tch-prob 42 Problem 95 on Page 83 95 Consider a well-shuffled deck of cards consisting of 52 distinct cards, of which four are aces and four are kings. a. Pr[Obtaining an ace in the first draw]= 4 1 52 13 b. Draw a card from the deck and look at it. Pr[Obtaining an ace in the second draw]= A 3 51 A 4 51 If does not know the previous outcome 1 13 4 513 12 13 51 1 13 tch-prob 43 c. Suppose we draw 7 cards from the deck. Pr[7 cards include 3 aces]= Pr[7 cards include 2 kings]= 4 48 3 4 52 7 4 48 2 5 52 7 Pr[7 cards includes 3 aces and 2 kings]= 4 4 44 3 2 2 52 7 d. Suppose the entire deck of cards is distributed equally among four players. Pr[each player gets an ace]= 4 48 3 36 2 24 13 1 12 1 12 1 12 13 4! 52 39 26 13 13 13 13 13 4! tch-prob 44 Birthday problem. p =No two people in a group of n people will have a common birthday. 1 2 n 1 p 1 1 365 1 365 365 1 p n 23 2 p 0.01 n 56 Lotto Problem. Six numbers and one special number are picked from a pool of 42 numbers. … 1. What is the probability of winning the grand prize? 2. What is the probability of winning 2nd prize? 3. What is the probability that number 37 is drawn next time? tch-prob 45