November 2002 Course 4 solutions

Similar documents
Estimation for Complete Data

Direction: This test is worth 150 points. You are required to complete this test within 55 minutes.

Simulation. Two Rule For Inverting A Distribution Function

Statistical Inference (Chapter 10) Statistical inference = learn about a population based on the information provided by a sample.

EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY

Direction: This test is worth 250 points. You are required to complete this test within 50 minutes.

This exam contains 19 pages (including this cover page) and 10 questions. A Formulae sheet is provided with the exam.

Review Questions, Chapters 8, 9. f(y) = 0, elsewhere. F (y) = f Y(1) = n ( e y/θ) n 1 1 θ e y/θ = n θ e yn

Last Lecture. Wald Test

Stat 319 Theory of Statistics (2) Exercises

x = Pr ( X (n) βx ) =

Let us give one more example of MLE. Example 3. The uniform distribution U[0, θ] on the interval [0, θ] has p.d.f.

1.010 Uncertainty in Engineering Fall 2008

STAT431 Review. X = n. n )

Unbiased Estimation. February 7-12, 2008

NANYANG TECHNOLOGICAL UNIVERSITY SYLLABUS FOR ENTRANCE EXAMINATION FOR INTERNATIONAL STUDENTS AO-LEVEL MATHEMATICS

Topic 9: Sampling Distributions of Estimators

Topic 9: Sampling Distributions of Estimators

Econ 325 Notes on Point Estimator and Confidence Interval 1 By Hiro Kasahara

Econ 325/327 Notes on Sample Mean, Sample Proportion, Central Limit Theorem, Chi-square Distribution, Student s t distribution 1.

Linear regression. Daniel Hsu (COMS 4771) (y i x T i β)2 2πσ. 2 2σ 2. 1 n. (x T i β y i ) 2. 1 ˆβ arg min. β R n d

Asymptotics. Hypothesis Testing UMP. Asymptotic Tests and p-values

Important Formulas. Expectation: E (X) = Σ [X P(X)] = n p q σ = n p q. P(X) = n! X1! X 2! X 3! X k! p X. Chapter 6 The Normal Distribution.

Mathematical Statistics - MS

FACULTY OF MATHEMATICAL STUDIES MATHEMATICS FOR PART I ENGINEERING. Lectures

STATISTICAL INFERENCE

Since X n /n P p, we know that X n (n. Xn (n X n ) Using the asymptotic result above to obtain an approximation for fixed n, we obtain

IIT JAM Mathematical Statistics (MS) 2006 SECTION A

Lecture 7: Properties of Random Samples

Discrete Mathematics for CS Spring 2008 David Wagner Note 22

of the matrix is =-85, so it is not positive definite. Thus, the first

Random Variables, Sampling and Estimation

Linear Regression Models

Let A and B be two events such that P (B) > 0, then P (A B) = P (B A) P (A)/P (B).

First Year Quantitative Comp Exam Spring, Part I - 203A. f X (x) = 0 otherwise

Solutions to Odd Numbered End of Chapter Exercises: Chapter 4

Section 14. Simple linear regression.

Probability and statistics: basic terms

SDS 321: Introduction to Probability and Statistics

TMA4245 Statistics. Corrected 30 May and 4 June Norwegian University of Science and Technology Department of Mathematical Sciences.

Goodness-of-Fit Tests and Categorical Data Analysis (Devore Chapter Fourteen)

Lecture 2: Monte Carlo Simulation

DS 100: Principles and Techniques of Data Science Date: April 13, Discussion #10

( θ. sup θ Θ f X (x θ) = L. sup Pr (Λ (X) < c) = α. x : Λ (x) = sup θ H 0. sup θ Θ f X (x θ) = ) < c. NH : θ 1 = θ 2 against AH : θ 1 θ 2

Open book and notes. 120 minutes. Cover page and six pages of exam. No calculators.

A Question. Output Analysis. Example. What Are We Doing Wrong? Result from throwing a die. Let X be the random variable

3/3/2014. CDS M Phil Econometrics. Types of Relationships. Types of Relationships. Types of Relationships. Vijayamohanan Pillai N.

7.1 Convergence of sequences of random variables

Summary. Recap ... Last Lecture. Summary. Theorem

EECS564 Estimation, Filtering, and Detection Hwk 2 Solns. Winter p θ (z) = (2θz + 1 θ), 0 z 1

Properties and Hypothesis Testing

The variance of a sum of independent variables is the sum of their variances, since covariances are zero. Therefore. V (xi )= n n 2 σ2 = σ2.

Lecture 11 and 12: Basic estimation theory

An Introduction to Randomized Algorithms

Estimation of the Mean and the ACVF

[ ] ( ) ( ) [ ] ( ) 1 [ ] [ ] Sums of Random Variables Y = a 1 X 1 + a 2 X 2 + +a n X n The expected value of Y is:

4. Partial Sums and the Central Limit Theorem

MATH 320: Probability and Statistics 9. Estimation and Testing of Parameters. Readings: Pruim, Chapter 4

January 25, 2017 INTRODUCTION TO MATHEMATICAL STATISTICS

Topic 9: Sampling Distributions of Estimators

TABLES AND FORMULAS FOR MOORE Basic Practice of Statistics

Chapter 6 Principles of Data Reduction

Solutions: Homework 3

AMS570 Lecture Notes #2

f(x i ; ) L(x; p) = i=1 To estimate the value of that maximizes L or equivalently ln L we will set =0, for i =1, 2,...,m p x i (1 p) 1 x i i=1

Basis for simulation techniques

Introduction to Econometrics (3 rd Updated Edition) Solutions to Odd- Numbered End- of- Chapter Exercises: Chapter 3

Kurskod: TAMS11 Provkod: TENB 21 March 2015, 14:00-18:00. English Version (no Swedish Version)

Problems from 9th edition of Probability and Statistical Inference by Hogg, Tanis and Zimmerman:

STAT Homework 1 - Solutions

Output Analysis (2, Chapters 10 &11 Law)

Agreement of CI and HT. Lecture 13 - Tests of Proportions. Example - Waiting Times

Lecture 12: September 27

Exponential Families and Bayesian Inference

Class 23. Daniel B. Rowe, Ph.D. Department of Mathematics, Statistics, and Computer Science. Marquette University MATH 1700

Power and Type II Error

Describing the Relation between Two Variables

Advanced Engineering Mathematics Exercises on Module 4: Probability and Statistics

Circle the single best answer for each multiple choice question. Your choice should be made clearly.

Math 152. Rumbos Fall Solutions to Review Problems for Exam #2. Number of Heads Frequency

Goodness-Of-Fit For The Generalized Exponential Distribution. Abstract

Statistical Intervals for a Single Sample

Questions and Answers on Maximum Likelihood

Lecture 33: Bootstrap

Statisticians use the word population to refer the total number of (potential) observations under consideration

Lecture 6 Chi Square Distribution (χ 2 ) and Least Squares Fitting

7.1 Convergence of sequences of random variables

Math 113, Calculus II Winter 2007 Final Exam Solutions

STAT 350 Handout 19 Sampling Distribution, Central Limit Theorem (6.6)

The Sampling Distribution of the Maximum. Likelihood Estimators for the Parameters of. Beta-Binomial Distribution

MATH 472 / SPRING 2013 ASSIGNMENT 2: DUE FEBRUARY 4 FINALIZED

STAC51: Categorical data Analysis

Lecture 6 Simple alternatives and the Neyman-Pearson lemma

Lecture 6 Chi Square Distribution (χ 2 ) and Least Squares Fitting

Assessment and Modeling of Forests. FR 4218 Spring Assignment 1 Solutions

Exam II Review. CEE 3710 November 15, /16/2017. EXAM II Friday, November 17, in class. Open book and open notes.

This section is optional.

CEE 522 Autumn Uncertainty Concepts for Geotechnical Engineering

Problem Set 2 Solutions

Department of Civil Engineering-I.I.T. Delhi CEL 899: Environmental Risk Assessment HW5 Solution

Transcription:

November Course 4 solutios Questio # Aswer: B φ ρ = = 5. φ φ ρ = φ + =. φ Solvig simultaeously gives: φ = 8. φ = 6. Questio # Aswer: C g = [(.45)] = [5.4] = 5; h= 5.4 5 =.4. ˆ π =.6 x +.4 x =.6(36) +.4(4) = 384..45 (5) (6) Questio # 3 Aswer: D N is distributed Poisso(λ) µ = E( λ) = αθ = (.) =.. v= E λ = a= Var λ = αθ = =. 5 k = = ; Z = =..44 6 + 5/ 6 7 Thus, the estimate for Year 3 is 5 (.5) + (.) =.4. 7 7 ( ).; ( ) (.).44. Note that a Bayesia approach produces the same aswer.

Questio # 4 Aswer: C At the time of the secod failure, ˆ 3 Ht () =. + = 3 = At the time of the fourth failure, ˆ Ht () = + + + =.3854. 9 Questio # 5 Aswer: E R ˆ x 4 = =.65 =.984. 8 i β yi Questio # 6 Aswer: B The likelihood is: x j rr ( + ) L( r+ x ) j β r+ x j j= x j!( + β) j= x r x j j L = β ( + β). The loglikelihood is: l = xjl β ( r+ xj)l( + β) j= xj r+ xj l = = j= β + β = x ( + β) ( r+ x ) β = x rβ j j j j= j= = x rβ; βˆ = x / r.

Questio # 7 Aswer: C The Bühlma credibility estimate is Zx+ ( Z ) µ where x is the first observatio. The Bühlma estimate is the least squares approximatio to the Bayesia estimate. Therefore, Z ad µ must be selected to miimize + µ + + µ + + µ 3 3 3 [ Z ( Z).5] [ Z ( Z).5] [3 Z ( Z) 3]. Settig partial derivatives equal to zero will give the values. However, it should be clear that µ is the average of the Bayesia estimates, that is, µ = (.5 +.5 + 3) =. 3 The derivative with respect to Z is (deletig the coefficiets of /3): ( Z +.5)( ) + (.5)() + ( Z )() = Z =.75. The aswer is.75() +.5() =.5. Questio # 8 Aswer: E / The cofidece iterval is ˆ θ ˆ θ ( St ( ), St ( ) ). Takig logarithms of both edpoits gives the two equatios l.695 =.36384 = l St ˆ( ) θ l.843 =.779 = θ l St ˆ( ). Multiplyig the two equatios gives.64 = [l St ˆ( )] l St ˆ( ) =.498 St ˆ( ) =.77936. The egative square root is required i order to make the aswer fall i the iterval (,).

Questio # 9 Aswer: E Because k ρk = φ there are a umber of ways to get the value of φ. / /3. φ =.5 =.46368; φ = (.) =.4646; φ = =.465..5 Also, because the mea is zero, δ must be zero. The (usig the first choice for φ), y ˆ + =.46368(.43) + =.998. T Questio # Aswer: B The likelihood is: α α α α5 α5 α5 L = (5 + 5) (5 + 55) (5 + 95) 3 3α α 5 =. α + (375i675i) The loglikelihood is: α+ α+ α+ l = 3lα + 3α l5 ( α + ) l(375i675i) 3 3 l = + 3l5 l(375i675i ) = 4.48 α α ˆ α = 3/ 4.48 =.6798. Questio # Aswer: D For this problem, r = 4 ad = 7. The, 33.6 v ˆ = =.4 4(7 ) The, ad 3.3.4 a ˆ = =.9. 4 7.4 4 7 63 k = = ; Z = = =.8..9 9 7 + (4 / 9) 77

Questio # Aswer: A r r r.6.5.4 YX YX 3 X X 3 YX. X = = =.54 3 ryx r.5.4 3 X X 3 r Questio # 3 Aswer: C For a mixture variable, raw momets are weighted averages of the idividual momets. Thus, EX ( ) =.5 m+.5m ad EX ( ) =.5( m) +.5( m). The square of the coefficiet of variatio is VarX ( ) EX ( ) EX ( ) m + m = =. EX ( ) EX ( ).5( m m) + Divide umerator ad deomiator by variatio becomes m ad let r m/ m =. The square of the coefficiet of r +. Settig the derivative equal to zero yields r =, however, this value miimizes.5( r + ) the fuctio (at a value of ). There are o other critical poits. Lookig at the edpoits (r = ad r = ifiity) the limitig value is 3, which is the maximum. Therefore, the least upper boud for the coefficiet of variatio is the square root of 3. Questio # 4 Aswer: B X is the radom sum Y + Y +... + Y N. N has a egative biomial distributio with r = α = 5. ad β = θ =.. ( ) = rβ =.3 ( ) rβ( β) E N Var N bg = + =.36 EY = 5 Var bg Y = 5,,

( ) =.3 5 = 5 E X Var( X ) =.3 5,, +.36 5,, = 6,5, Number of exposures (isureds) required for full credibility ( ) FULL = (.645/.5) 6,5, / 5 = 7937.67. Number of expected claims required for full credibility EN ( ) FULL =.3 7937.67 = 38. Questio # 5 Aswer: C The estimated relative risk is b ht ( Z= ) h () t e b = = e =.8 b=.6. ht ( Z= ) h() t For the sigle covariate case, the Wald test for testig H : β = reduces to: ( b ) I( b) = (.6) (3.968) =.43. Questio # 6 Aswer: D See pages 535-7, the bottom of page 567 ad the top of page 568. The oly correct statemet ad the correct aswer is (D). Questio # 7 Aswer: E X Fb xg d i ( ) F x b g b g F x F x F x F x F x d i 9..5.5.5 64.4..473.73.73 9.6.4.593.7.93 35.8.6.74.59.4 8..8.838.6.38 where: ˆ θ = x = ad ( ) x / F x = e. The maximum value from the last two colums is.73. bg

Questio # 8 Aswer: E µ = E λ = v= E σ = a= Var λ = k = v/ a= 5; Z = =. + 5 6 ( ) ; ( ).5; ( ) /. Thus, the estimate for Year is 5 () + () =.9375. 6 6 Questio # 9 Aswer: B Time must be reversed, so let T be the time betwee accidet ad claim report ad let R = 3 T. The desired probability is Pr( T < T 3) = Pr(3 R< 3 R 3) = Pr( R> R ) = Pr( R> ). The product-limit calculatio is: R Y d 4 9 55 3 43 43 The estimate of survivig past (reversed) time is (3/4)(3/55) =.459. Questio # Aswer: E The solutio depeds o idetifyig the parameter a with the overall average, which requires the covetio stated i Pidyck ad Rubifeld uder the table o page 4. The covetio also applies to subgroups. With this covetio, the solutio is the followig coditioal expectatio: (,,, ) E Y E = F = G = H = a= b b c c.

Questio # Aswer: E The posterior desity, give a observatio of 3 is: θ 3 f (3 θπθ ) ( ) (3 + θ) θ πθ ( 3) = = θ πθ θ θ θ 3 f (3 ) ( ) d (3 + ) d 3 (3 + θ ) = = + > (3 + θ ) 3 3(3 θ), θ. The, 6 5 3 Pr( Θ> ) = 3(3 + θ) dθ = 6(3 + θ) = =.64. Questio # Aswer: B Because all previous values are, previous sigle ad double smoothed values are also zero. The, y =.6() +.4() =.6 y =.6(.) +.4(.6) =.96 y =.6(.3) +.4(.96) =.64 y =.6(.6) +.4() =.36 y =.6(.96) +.4(.36) =.7 y =.6(.64) +.4(.7) =.9864. Questio # 3 Aswer: B L= F() [ F() F()] [ F()] 7 6 7 = ( e ) ( e e ) ( e ) / θ 7 / θ / θ 6 / θ 7 = ( p) ( p p ) ( p ) = p ( p) 7 6 7 3 where p θ = = / θ = e. The maximum occurs at p = /33 ad so ˆ / l( / 33) 996.9.

Questio # 4 Aswer: A EXθ ( ) = θ /. 6 3(6 ) EX ( 3 4,6) EX ( ) f( 4,6) d θ θ = θ θ θ = 3 dθ 6 = 6 4 θ 3 3(6 )(6 ) = = 45. 4 3 3 6 Questio # 5 Aswer: D The data may be orgaized as follows: t Y d St ˆ( ) (9/) =.9 3 9.9(7/9) =.7 5 7.7(6/7) =.6 6 5.6(4/5) =.48 7 4.48(3/4) =.36 9.36(/) =.8 Because the product-limit estimate is costat betwee observatios, the value of S ˆ(8) is foud from S ˆ(7) =.36. Questio # 6 Aswer: D As of time there were 7 observed paymets, so O = 7. For the Weibull distributio, the cumulative hazard fuctio is H( x) = l S( x) = x /5. The xi EZ ( ) = VarZ ( ) = = (4 + 9 + 9 + 5 + 5 + 36 + 49 + 49 + 8+ ) = 5.48. i= 5 5 The chi-squared test statistic (with oe degree of freedom) is (7 5.48) /5.48 = 4.645. From the tables, this leads to rejectio at the 5% level, but ot at the.5% level.

Questio # 7 Aswer: D ESSR = 5, 5,565 = 9, 435 ESSUR RUR =.38 = ESSUR =.6(5,) = 9,3 TSS (9, 435 9,3) / 3 F = = 5.7. 9,3 / 3,4 Questio # 8 Aswer: C The maximum likelihood estimate for the Poisso distributio is the sample mea: ˆ 5() + () + () + 9(3) λ = x = = 365.6438. The table for the chi-square test is: Number of days Probability Expected* Chi-square.6438 e =.934 7.53 5.98.6438.6438 e =.3765 5.94.3.6438.6438 e 95.3.34 =.68 3+.83** 83.3.9 *365x(Probability) **obtaied by subtractig the other probabilities from The sum of the last colum is the test statistic of 7.56. Usig degrees of freedom (4 rows less estimated parameter less ) the model is rejected at the.5% sigificace level but ot at the % sigificace level.

Questio # 9 Aswer: D.4() +.() +.().() +.() +.() µ () = =.5; µ () = =.6.4 µ =.6(.5) +.4() =.7 a =.6(.5 ) +.4( ).7 =.6.4() +.() +.(4) 7.() +.() +.(4) v.6.4 v =.6(7 /) +.4(.5) = / 6 k = v/ a= 55/6; Z = = + 55/ 6 5 () =.5 = ; v() = =.5 Bühlma credibility premium = 6 + 55 (.7) =.8565. 5 5 Questio # 3 Aswer: A All the statemets about R are true, but oly (A) is ot raised as a objectio to R. Questio # 3 Aswer: C µ =.5() +.3() +.() +.(3) =.8 σ =.5() +.3() +.(4) +.(9).64 =.96 3 ES ( ) = σ = (.96) =.7 4 bias =.7.96 =.4.

Questio # 3 Aswer: C The four classes have meas.,.,.5, ad.9 respectively ad variaces.9,.6,.5, ad.9 respectively. The, µ =.5(. +. +.5 +.9) =.45 v =.5(.9 +.6 +.5 +.9) =.475 a =.5(. +.4 +.5 +.8).45 =.96875 k =.475 /.96875 =.558 4 Z = =.743 4 +.558 The estimate is [.743( / 4) +.757(.45)] 5 =.4. Questio # 33 Aswer: D The lower limit is determied as the smallest value such that St ˆ( ).5 +.96 VSt ˆ[ ˆ( )]. At t = 5 the two sides are.36 ad.5 +.96(.47) =.34 ad the iequality does ot hold. At t = 54 the two sides are.93 ad.5 +.96(.456) =.339 ad the iequality does hold. The lower limit is 54. Questio # 34 Aswer: A (( ) ) K ˆk... 8.96 k = Q= T r = + + + = Questio # 35 Aswer: A The distributio used for simulatio is give by the observed values.

Questio # 36 Aswer: B First obtai the distributio of aggregate losses: Value Probability /5 5 (3/5)(/3) = /5 (/5)(/3)(/3) = 4/45 5 (3/5)(/3) = /5 5 (/5)()(/3)(/3) = 4/45 4 (/5)(/3)(/3) = /45 µ = (/ 5)() + (/ 5)(5) + (4 / 45)() + ( / 5)(5) + (4 / 45)(5) + (/ 45)(4) = 5 σ = (/ 5)( ) + (/ 5)(5 ) + (4 / 45)( ) + ( / 5)(5 ) + (4 / 45)(5 ) + (/ 45)(4 ) 5 = 8,. Questio # 37 Aswer: A Loss Rage Cum. Prob..3.53 4.8 4 75.96 75.98 5. At 4, F(x) =.8 = e 4 θ ; solvig gives θ = 48.53. Questio # 38 Aswer: D The sum of the squared values of the idepedet variable is 66 + () =,66. The value of s is 5348/8 = 97.. The, ΣX i, 66 sˆ ˆ α = s = 97. = 36. N x Σ i (66) The 97.5 th percetile of a t-distributio with 8 degrees of freedom is., so the lower limit of the symmetric 95% cofidece iterval for α is 68.73. 36 = 7.78.

Questio # 39 Aswer: B (/ )(/ 3) 3 Pr( class ) = = (/ )(/ 3) + (/ 3)(/ 6) + (/ 6)() 4 (/ 3)(/ 6) Pr( class ) = = (/ )(/ 3) + (/ 3)(/ 6) + (/ 6)() 4 (/ 6)() Pr( class3 ) = = (/ )(/ 3) + (/ 3)(/ 6) + (/ 6)() because the prior probabilities for the three classes are /, /3, ad /6 respectively. The class meas are µ () = (/3)() + (/3)() + (/3)() = µ () = (/ 6)() + (/3)() + (/ 6)(3) =. The expectatio is EX ( ) = (3/ 4)() + (/ 4)() =.5. Questio # 4 Aswer: E The first, secod, third, ad sixth paymets were observed at their actual value ad each cotributes f(x) to the likelihood fuctio. The fourth ad fifth paymets were paid at the policy limit ad each cotributes F(x) to the likelihood fuctio. This is aswer (E).