You are on page 1of 14
Tata McGraw-Hill NA | Probability, Statistics and Random Processes ay Veerarajan Information contained in this work has been obtained by Tata McGraw-Hill, from sources believed to be reliable. However, neither Tata McGraw-Hill nor its authors guarantee the accuracy or completeness of any information published herein, and neither Tata McGraw-Hill nor its authors shall be responsible for any errors, omissions, or damages arising out of use of this information. This work is published with the understanding that Tata McGraw-Hill and its authors are supplying information but are not attempting to render engincering or other professional services. If such services are required, the assistance of an appropriate professional should be sought. Tata McGraw-Hill © 2003, 2002, Tata McGraw-Hill Publishing Company Limited Eleventh reprint 2007 RALLCRXYRCLYX No part of this publication can be reproduced in any form or by any means without the prior written permission of the publishers This edition can be exported from India only by the publishers, Tata McGraw-Hill Publishing Company Limited ISBN 0-07-049482-7 Published by Tata McGraw-Hill Publishing Company Limited, 7 West Patel Nagar, New Delhi 110 008, typeset at Script Makers, 19 A-1-B DDA Market, Paschim Vihar, New Delhi 110 063 and printed at Rashtriya Printers, M-135 Panchsheel Garden, Naveen Shahdara, Delhi 110032 Contents Preface to the Second Edition v Preface to the First Edition vii 1. Probability Theory 1 Random Experiment _/ Mathematical or Apriori Definition of Probability J Statistical or Aposteriori Definition of Probability 2 Axiomatic Definition of Probability 2 Conditional Probability 4 Independent Events 5 Theorem of Total Probability _/9 Baye’s Theorem or Theorem of Probability of Causes /9 B i's Tri 25 De Moivre—-Laplace Approximation 26 G lisation of B Ll’ Th Multinomial Distributi 2 2. Random Variables 36 Discrete Random Variable 36. Probability Function 37 Continuous Random Variable _37 Probability Density Function 37 Cumulative Distribution Function (cdf) 38 Properties of the cdf F(x) 38 Special Distributions 38 Probability Function of (X,Y) 59 Joint Probability Density Function 59 Properties of F(x, 60 Marginal Probability Distribution 60 x Contents Conditional Probability Distribution _6/ Independent RVs_6/ Random Vectors 62 Marginal Probability Distribution of X: {i, p,*} 66 Marginal Probability Distribution of Y: {j,p*;} 66 3. Funeti { Random Variabl 89 Function of Onc Random Variable 89 One Function of Two Random Variables 90 Two Functi fTwoR Vari 93 4. Statistical Averages 120 Expected Values of a Two-Dimensional RV /21 Properties of Expected Values 122 Covari C . 23 Properties 123 Conditional Expected Values 124 Properties /25 Properties of MGF 14] Properties of Characteristic Function _/42 Comulant Generating Function (CGF) _/44 Bounds on Probabilities 155 Tchebycheff Inequality /55 Bienayme’s Inequality 156 Schwartz Inequality 157 Cauchy-Schwartz Inequality 157 Convergence Concepts and Central Limit Theorem _/64 Central Limit Theorem (Liapounoff’s Form) /65 Central Limit Theorem (Lindberg-Levy’s Form) _/65 Some Special Probability Distributions 178 Introduction _178 Special Discrete Distributions 178 Mean and Variance of the Binomial Distribution 179 Recurrence Formula for the Central Moments of Poisson Distribution as Limiting Form of Binomial Distribution /8/ Mean and Variance of Poisson Distribution 183 Recurrence Formula for the Central Moments of the Poisson Distribution 183 Mean and Variance of Hypergeometric Distribution 186 ye 6. Contents: xi Binomial Distribution as Limiting Form of Hypergeometric Distribution 188 Special Continuous Distributions 2/0 Moments of the Uniform Distribution U (a,b) 2/1 Mean and Variance of the Exponential Distribution 2/2 Memoryless Property of the Exponential Distribution 2/3 Mean and Variance of Erlang Distribution 2/4 Reproductive Property of Gamma Distribution 2/4 Relation between the Distribution Functions (cdf ) of the Erlang Distribution with 4 = 1 (or Simple Gamma Distribution) and (Poisson Distribution) 2/5 Density Function of the Weibull Distribution 2/6 M Vari € Weibull Distribution 216 Standard Normal Distribution 2/7 Normal Probability Curve 2/8 Properties of the Normal Distribution N (uo) 2/8 Importance of Normal Distribution _227 Tests of Hypotheses 263 P. Statistics 262 Sampling Distribution 263 Estimation and Testing of Hypotheses 264 Tests of Hypotheses and Tests of Significance 264 Critical Region and Level of Significance 265 Errors in Testing of Hypotheses 266 Critical Values or Significant Values 266 Procedure for Testing of Hypothesis 267 Interval Estimation of Population Parameters 268 Tests of Significance for Large Samples 268 Tests of Significance for Small Samples 291 S Te 1-Distributi 29 Properties of t-Distribution 292 U f¢-Distributi 202 Critical \ f Table 203 s *¢ F-Distributi 205 Properties of the F-Distribution 296 Use of F-Distributi 206 F-test 296 Chi-Square Distribution 341 Properties of y?-Distribution 3/7 Uses of y7-Distribution 311 2°-Test of Goodness of Fit 3/2 Conditions for the Validity of x * Test 312 X°-Testof Independence of Attributes 3/2 Random Variables 63 L(x1+%2543) fxs x(%245) ———___———|] Worxep Exampte 2(8) |} ———_____ Examp.e 1 Three balls are drawn at random without replacement from a box containing 2 white, 3 red and 4 black balls, If X denotes the number of white balls drawn and Y denotes the number of red balls drawn, find the joint probability distribution of (x, /. As there are only 2 white balls in the box, X can take the values 0, 1 and 2 and Y can take the values 0, 1, 2 and 3. P(X=0, Y=0) = P(drawing 3 balls none of which is white or red) = P(all the 3 balls drawn are black) Se, fg x5) = 1 =4CY9C3= P(X=0,¥=1) = P(drawing | red and 2 black balls) 3C,x4C, _ 3 “~oc, 14 Similarly, P(X =0, ¥=2)=3@2%4G 21; pao, y=3)2b 9G. 7 84 1 2 1 PX = 1, ¥=0)= 55 PX= 1, Y= l= 5) PXa 1 Ya Va Ts P(X=1,Ys = (since only 3 balls wre tevin) P(X= 2Y=H= 4; P(X=2, Y= De ts Pes 2, ¥=2)=0; P(X=2,¥=3)= 3 The joint probability distribution of (X, Y) may be represented in the form of a table as given below: xX Y 0 L 2 3 o jt] a2) 2 42 21 14 7 84 , | a | ie | 7 1 2 1 t 7 | 7 | @ | ° 2 a + 0 0 21 28 64 Probability, Statistics and Random Processes Sum all the cell probabilities EXAMPLE 2 For the bivariate probability distribution of (X, Y) given below, find P(X < 1), P(YS3), P(X $1, YS 3), P(X S W/¥ $3), P(Y S 3/X S 1) and P(X + YS 4). PX < 1) = P(X =0) + P(X =1) 6 6 =} P(xX=0,¥=f)+ ¥ P(k=1,¥=/) jst jal =(o+0 ort e324 2y A)e(hahepedeies) 32°32 30732 16 16 8 8 8 8 1,5_7 sotiee 4 8 8 P(Y Ss 3) = P(Y = 1) + P(Y = 2) + P(¥ =3) -> P(X =i, ¥=1)+ s P(X =i, Y= 2) i=0 izo 2 +> P(X=i, ¥=3) 1 1 1 1 1 1.1 =(0+—+— +—4+—]+(—4+i4+— (orga) +(° 16 3) ca 8 zi) 3,3, 23 s+ “32°32 64 64 3 3 PIXS1,¥S3)= ¥ P(X=0,¥=)+ ¥ Pk=LY=/ jel jal Random Variables 65 (o+0 +t) (4+4+3) = 2 32, 16 16 8) 32 P(XS1,YS3) _ 9/32 _ 18 P(X S VY S3)= = = ‘ = SY s3) 23/64 23 P(XSL,YS3) 9/32 _ 9 PY S3/X $1) = = = ‘ ) P(X S1) 718 28 2 ial 3 PUX+¥S4)= y PUX=0,¥=f+ Y PX=1,Y=j+ jel j=l 113 =241h B 32° 4 16 32 EXAMPLE 3 The joint probability mass function of (X, Y) is given by p(x, y) = k(2x + 3y), x=0, 1,2;y=1,2,3. Find all the marginal and conditional probability distribu- tions. Also find the probability distribution of (X + ¥). The joint probability distribution of (X, Y) is given below. The relevant prob- abilities have been computed by using the given law. ¥ x euuiqti L 2 3 0 3k 6k 9k 1 5k 8k 11k 2 Tk 10k 13k 3.2 LE ply y=1 ga1iFo0 ie., the sum of all the probabilities in the table is equal to 1. ie, 72k = 1, Random Variables 79 20. 21. . ffx, y) =k -x-y),0 ¥) and P{max (X, Y) = 3) and Gi) Find the probability distribution of the RV Z = min (X, ¥). The input to a binary communication system, denoted by a RV X, takes on one of two values O or 1 with probabilities 3/4 and 1/4 respectively. Because of errors caused by noise in the system, the output Y differs from the input eccasionally. The behaviour of the communication system is modeled by the conditional probabilities given below: P(Y = WX = 1) = 3/4 and P(Y= 0/X = 0) = 7/8 Find (i) P(Y = 1), (ii) PCY = 0) and (iii) P(X = 1/Y = 1). The following table represents the joint probability distribution of the discrete RV (X, Y). Find all the marginal and conditional distributions. 184 Probability, Statistics and Random Processes Differentiating (1) with respect to 2, we get Ste SF ay et ae ret ate ay (rN qa Ayr! =k py + xy Jagr (rae rao 1! 1 SOKA t > Hee ten Hin =A( SHE + kan, ) 2 ‘Using recurrence relation (2), we may compute moments of higher order, provided we known moment of lower order. Putting k = 1 in (2), we get du i= (2H + uy | = Ao uy =1 andy, = 0) Puttingk = 2 in (2), we get due wy=a (S82 4 2,)= Putting k= 3 in (2), we get =a (2B «3 | = 208+ 1 da The interesting property of the Poisson distribution is the equality of its mean, variance and third-order central moment. 3. Geometric distribution Definition: Let the RV X denote the number of trials of a random experiment required to obtain the first success (occurrence of an event A). Obviously X can assume the values 1, 2, 3, ... Now X =r, if and only if the first (r ~ 1) trials result in failure (occurrence of A) and the rth trial results in success (occurrence of A). Hence PX =n=q |p, r=1,23,.0 where P(A) =p and P(A) =4. If X is a discrete RV that can assume the values 1, 2, 3, probability mass function is given by P(X =r) =q"' p; r=1,2,....0 where p+qg=1 then X is said to follow a geometric distribution, » e@ such that its Some Special Probability Distributions 185. Geometric distribution is a legitimate probability distribution, since L PK = Dg 'p rst ret =plitqt¢ +...) > = =1 1-4 Mean and Variance of Geometric Distribution EX) = 3 x,p, x rap ret =pll+2q+3¢'+...+ 0] =p(l-grt=4 P E(X*) = x2 pr ; = 3 Pq"p rai =pd (re -r} q™ rel =pi{(lx2+ 2x3g43xX4Q'4...+00) = (1+29q4+3q7+...+0)] = pla =)? --9)"l L 1 => Q-p=s (+ p & P) pS ? 1 1 = (l4+q)-=-4. P pp Sometimes the probability mass function of a geometric RV X is taken as PIX =r) =" prr=O,1,2,...,0 whe ptget Isis this definition that was given in chapter IL. If this definition is assumed, then E(x) = © and Var(X) = Le [see example (5) in section 4 Vv, Pp v 186 Probability, Statistics and Random Processes 4. Hypergeometric distribution 1fX represents the number of defectives found, when n items are drawn without replacement from a lot of N items containing k defectives and (N — k) non- defectives, clearly KC,-(N-K)G,_ — , P(X r= We n =0, 1,2, ..., min (n, k) Ifn > &, then the maximum value of X és kz If n &, can take the values 0, 1, 2, ...k In other words, PIX =n) =0, when r= b+ 1, b+ 2, ....n, This value (namely zero) of the probability is provided by the probability mass function fornula itself. since RC. = 0, for r=kh+lkt+2, 040 Thus in the walue of P(X = r) min (n, &) can be replaced by n. (2) Hyperpeometric distribution is a legitimatte probability distribution, since a £6, (N-AC yy) PX =n = — 1, x esa z NC, ” =—_ NG, =1 sine NC, * DX AC, (N- 4) C,., = coefficient of x" in (1 + xf* (1 + x\-* r20 coefficient of x" in (1 + 3) NG, Mean and Variance of Hypergeometric Distribution EX) =D x7, " =D rk, -(V- W)Cy,,INC, r=0 Some Special Probability Distributions 187 (k=-1) C N-DC, “aed (k= 1) Cy NK) Conny Ee, 1k as oy (on putting K = k- Landy =r-1) (N - 1)C, _ 1, [by step (1) in note (2) given above] atk “7 E(X?) = E(X(X - 1) + X) n = E(X(K—D} + = nk. & = 72 SY rr- I kC,-(N- WC _ yf NG, r=0 nk k(k-1) 2 = tw g (k= 2)C ip 2) W- WC, ak N k(k-1) "— es KCp (N=2= KC ya (on putting # =k—2 and’ =r—2) WONG, [by step (1) in note (2) given above] uk, k(k-Dn(n-1) N N(N-1) Var(X) = E(X?) — (E(X)¥? k(k—Iea(n—1) 2 N(N-1) NP “WN [MW = 1) + N= 1) (a= 1) - (NV - Dnk] =— tk _ N?(N-1) nk Fwy (V? — Nn — Nk + nk] _ nk(N—k)(N-n) ~~ NPN = 1) 280 Probability, Statistics and Random Processes EXamMPLe 10 The mean breaking strength of the cables supplied by a manufacturer is 1800, with an SD of 100. By a new technique in the manufacturing process, it is claimed that the breaking strength of the cable has increased. To test this claim, a sample of 50 cables is tested and it is found that the mean breaking strength is 1850. Can we support the claim at 1% LOS? Here X = 1850, n = 50, w= 1800 and o = 100. Hy ¥ =p. Ay x > One-tailed (right-tailed) test is to be used. Let LOS. = 1%. Therefore, z,,= 2.33. FH _ 1850-1800 _, 5, olin 100/150 Izleze Therefore, the difference between X and jis significant at 1% level, i. is rejected and H, is accepted. That is, based on the sample data, we may support the claim of increase in breaking strength. is Hy Exampce 11 The mean value of a random sample of 60 items was found to be 145, with an SD of 40, Find the 95% confidence limits for the population mean. What size of the sample is required to estimate the population mean within 5 of its actual value with 95% or more confidence, using the sample mean? Since the population SD is not given, we can approximate it by the sample SDs. Therefore, 95% confidence limits for 4 are given by val £1.96, sijn ie. ¥=-196 sp <¥ 41.96 + Le. x Va ru x + Ta ie. 14s — £9640 1.96 x40 Vo su S15 + To ie. 134.9

You might also like