25 views

Uploaded by Danar Handoyo

Statistical Modeling in Non-life Insurance.
Basic Bayesian analysis: fundamental concepts with examples, historical connection between Bayesian statistics and actuarial science, basic sampling scheme- inversion method, rejection sampling,
importance sampling and so on

Statistical Modeling in Non-life Insurance.
Basic Bayesian analysis: fundamental concepts with examples, historical connection between Bayesian statistics and actuarial science, basic sampling scheme- inversion method, rejection sampling,
importance sampling and so on

Attribution Non-Commercial (BY-NC)

- Sport and Participation Levels
- Bayesian
- Pharmacodynamic Analysis of a Fluid Challenge :: Critical Care 2016
- Bayes Decision Theory 2012
- Week #7 Homework Answers
- Methods of Animal Population Estimation
- UT Dallas Syllabus for stat7330.0u1.08u taught by Michael Baron (mbaron)
- six sigma
- ultima presentacion BayesianCourse
- 2007- Determinants of Growth Will Data Tell
- cogsysII-9.pdf
- Smith 1977
- IMChap007.docx
- Paper Gamma
- rpra5
- Use of Transformation When Comparing Two Means
- Jawaban UTS
- How to Create Custom Substitution Models _ BEAST Documentation
- Wright
- Technical Summaries of Scotian Slope

You are on page 1of 20

Statistical Research CNA Insurance Company Yanwei.Zhang@cna.com http://www.actuaryzhang.com

Bayesian statistics

1 / 20

Outline

Todays agenda

Bayesian analysis as a measure of uncertainty

Probability and Bayes Theorem Frequentists condence intervals

Bayesian statistics

2 / 20

What is probability

1

Frequency interpretation

Limiting proportion of times the event occurs in an innite sequence of independent repetitions of the experiment Examples: toss of coins, dices The requirement that an experiment can be repeated is rather limiting: does it make sense to make the statement of the form probability of rain tomorrow =0.5? What does it mean for independent experiment? (circular use of probability)

Subjective probability

Concern the assessments of a given person about uncertain outcomes Interpreted as a personal belief or a statement about uncertainty Based on the present knowledge of the event - knowledge base change leads to probability updating

Bayesian statistics

3 / 20

The inference process in a Bayesian analysis proceeds as follows: Specify a prior distribution p (), which reects our knowledge or uncertainty about prior to observing the data D ; Specify a probability model p (D |); Compute a posterior distribution p (|D ) for using the Bayes Theorem, which reects our knowledge of after observing the data.

Liklihood Posterior Prior

Bayesian statistics

4 / 20

Example

You are given:

1

A portfolio of independent risks is divided into two classes, Class A and Class B. There are twice as many risks in Class A as in Class B. The number of claims for each insured during a single year follows a Bernoulli distribution. Classes A and B have claim size distributions as follows:

Claim Size 50,000 100,000 Class A 0.60 0.40 Class B 0.36 0.64

2 3

The expected number of claims per year is 0.22 for Class A and 0.11 for Class B.

One insured is chosen at random. The insureds loss for two years combined is 100,000. You are asked to determine which class the selected insured is most likely to belong to.

Wayne Zhang (CNA insurance company) Bayesian statistics June 6, 2011 Chicago 5 / 20

Solution

Denote D as the event that the two years combined loss is 100,000, and in this case {A, B } . Prior distribution: p ( = A) = 0.6667, p ( = B ) = 0.3333. Data distribution: p (D |A) = 0.222 0.62 + 2 0.22 (1 0.22) 0.4 = 0.1547 p (D |B ) = 0.112 0.362 + 2 0.11 (1 0.11) 0.64 = 0.1269 Posterior distribution: P (A|D ) = p (D |A)P (A) p (D |A)P (A) + p (D |B )P (B ) 0.1547 0.6667 = 0.1547 0.6667 + 0.1269 0.3333 = 0.71

P (B |D ) = 1 P (A|D ) = 0.29

Wayne Zhang (CNA insurance company) Bayesian statistics June 6, 2011 Chicago 6 / 20

What do we learn?

It illustrates how data are combined with prior information to come to a conclusion The conclusion includes a measure of uncertainty Were we to use any other method, say MLE (choose that maximize p (D |)), we will not be able to get an estimate of uncertainty In that case, getting a condence interval is also meaningless: it suggests either complete condence or no condence In fact, the frequentists construction of condence intervals encounter many diculties.

Bayesian statistics

7 / 20

Condence interval

Consider the following example from Berger (1985): Suppose that X1 and X2 are independent with identical distribution given by P (Xi = 1) = P (Xi = + 1) = 1/2. Then, a frequentist 75% condence procedure is: (i.e., P ( (X ) = ) = 0.75 for all ) (X ) = (X1 + X2 )/2, X1 = X2 X1 1, X1 = X2

If x1 = x2 , we know (x1 + x2 )/2 = , thus (x ) is 100% certain to contain ; If x1 = x2 , (x ) is 50% certain to contain ; Neither is 75%! Whats going wrong?

Bayesian statistics

8 / 20

What the frequentists interval means is that in a sequence of repeated experiments, where each (x ) is constructed based on the experiment outcomes as above, the condence intervals will cover the true about 75% of the time. It is true with regard to overall performance because half the time the actual performance will be 100% and half the time the actual performance will be 50%. However, it clearly makes little sense to conduct an experiment, use (x ), and actually report 75% as the measure of accuracy!

Bayesian statistics

9 / 20

Prior information

1

A lady, who adds milk to her tea, claims to be able to tell whether the tea or the milk was poured into the cup rst. In all of ten trials conducted to test this, she correctly determines which was poured rst. A music expert claims to be able to distinguish a page of Haydn score from a page of Mozart score. In ten trials conducted to test this, he makes a correct determination each time. A drunken friend says he can predict the outcome of a ip of a fair coin. In ten trials conducted to test this, he is correct each time.

In all three situations, the unknown quantity is the probability of the person answering correctly. A classical signicance test of the various claims would reject the hypothesis that = 0.5 (i.e., the person is guessing) with a (one-tailed) signicance level of 210 .

Bayesian statistics

10 / 20

In situation 2 we would have no reason to doubt this conclusion. (The outcome is quite plausible with respect to our prior beliefs.) In situation 3, however, our prior opinion that this prediction is impossible (barring a belief in extrasensory perception) would tend to cause us to ignore the experimental evidence as being a lucky streak. In situation 1 it is not quite clear what to think, and dierent people will draw dierent conclusions according to their prior beliefs of the plausibility of the claim. In these three identical statistical situations, prior information clearly cannot be ignored.

Bayesian statistics

11 / 20

Subjective: the prior is chosen by making a serious attempt to quantify a state of knowledge prior to observing the data. E.g., I expect the loss ratio to follow log N (log (0.6), 0.32 ). Objective: prior is chosen to satisfy some objective criterion.

Noninformative prior: a prior that favors no particular values of the parameter over others, e.g., U (0, 2), N (0, 1002 ). Improper prior: a prior whose integral is innite, but the resultant posterior is a proper probability distribution, e.g., p () 1

In many cases, noninformative proper prior is in eect informative in the sense of favoring some parameter values It may lead to procedures with attractive frequentist properties

Jereys prior: motivated by the desire that inference should not depend on how a model is parameterized.

E.g., if U [0, 1], has a non-uniform distribution with higher density near 1 than 0. Obtained as p () I (), I () being the Fisher information

Bayesian statistics

12 / 20

Conjugate priors

Denition If F is a class of sampling distribution p (D |), and P is a class of prior distributions p (), then the class P is conjugate for F if p ( |D ) P p (D |) F , p () P

data distribution Binomial Poisson Exponential Normal ( known) conjugate prior Beta(, ) Gamma(, ) Gamma(, )

2 N ( 0 , 0 )

0 2 0

xi 2

v, v

v=

1 2 0

n 2

Bayesian statistics

13 / 20

Exchangeability

Exchangeability

Denition The random quantities x1 , , xn are said to be nitely exchangeable if P (x1 E1 , , xn En ) = P (x(1) E1 , , x(n) En ), for any permutation on the set 1, , n, and any sets E1 , , En of possible values. Intuitively, it says distributions of x1 , , xn do not depend on the labeling or order independent and identically distributed is a special case This formalizes the notion of the future being predictable on the basis of past experience. An innite sequence x1 , x2 , is exchangeable if every nite sequence is exchangeable

Wayne Zhang (CNA insurance company) Bayesian statistics June 6, 2011 Chicago 14 / 20

(1)

Exchangeability

Exchangeability contd

de Finettis Theorem If x1 , x2 , is an innite exchangeable sequence of random variables with probability measure P , there exists a probability measure Q on F , the set of all distributions on R, such that the joint distribution of x1 , , xn has the form

n

P ( x1 , , xn ) =

F i =1

F (xi )dQ (F )

(2)

1 2

Bayesian statistics

15 / 20

The posterior p (|D ) encapsulates the beliefs post observing the data, and various summaries are available Point estimate

Mean, median and mode Measure of spread such as variance, interquartile, quantiles

Interval estimate

100(1 )% credible interval C such that Highest probability density region (HPD): p (1 ) p (2 ), 1 C , 2 /C

C

p ()d = 1 .

Bayesian statistics

16 / 20

Example

Suppose that an insurance policy incurred 5 claims each year in a period of two years (n = 2): We assume the claim count follows a Poisson distribution yi | Pois () The prior comes from a conjugate Gamma distribution such that Gamma(, ) Then we have the posterior |y Gamma( + yi , + n). We rely on posterior simulation to draw inference, as this is more straightforward and can work when the posterior is not of closed form.

mean 3 5 8 8 8 Prior variation 1 3 4 20 100 mean 3.79 4.99 5.56 5.05 5.03 Posterior median 50% interval 3.72 [3.16, 4.37] 4.79 [3.98, 5.83] 5.42 [4.50, 6.47] 4.93 [3.99, 5.97] 4.84 [3.89, 5.98]

June 6, 2011 Chicago 17 / 20

Bayesian statistics

prior

posteior

mean=3, std=1 0

10

12

14

10

12

14

prior

posteior

mean=8, std=4 0

10

12

14

10

12

14

prior

posteior

mean=8, std=100 0

10

12

14

10

12

14

n= 5 n= 2

10 n= 20

12

10

12

n= 100

10

12

10

12

Bayesian statistics

19 / 20

Nonconjugate prior

Now suppose I want to use a Uniform prior that log N (0, 1002 ), then the posterior distribution will be |y

yi 1

exp

(log )2 2 1002

(3)

This is not a distribution (as a function of ) we are familiar with, so how do we simulate samples from it?

Bayesian statistics

20 / 20

- Sport and Participation LevelsUploaded bySPARC NZ
- BayesianUploaded byroninborges
- Pharmacodynamic Analysis of a Fluid Challenge :: Critical Care 2016Uploaded byJonathan Suarez
- Bayes Decision Theory 2012Uploaded bylcm3766l
- Week #7 Homework AnswersUploaded byAnonymous Qed2LOxF
- Methods of Animal Population EstimationUploaded byJohn Rudolf Catalan
- UT Dallas Syllabus for stat7330.0u1.08u taught by Michael Baron (mbaron)Uploaded byUT Dallas Provost's Technology Group
- six sigmaUploaded byShruti Gaind
- ultima presentacion BayesianCourseUploaded byyosoyanibal
- 2007- Determinants of Growth Will Data TellUploaded bySurjadi
- cogsysII-9.pdfUploaded byabhijeet borkar
- Smith 1977Uploaded bypeyman
- IMChap007.docxUploaded byEmarc Sagrado
- Paper GammaUploaded byG Nathan Jd
- rpra5Uploaded bymihai37
- Use of Transformation When Comparing Two MeansUploaded byLakshmi Seth
- Jawaban UTSUploaded byAndriyanto
- How to Create Custom Substitution Models _ BEAST DocumentationUploaded byThiago Accioly
- WrightUploaded byDarkyben
- Technical Summaries of Scotian SlopeUploaded byZulfani Aziz
- 09093011260425828Uploaded bymahoutsukai_monika
- Probability Distributrions of the Aggregated Residential LoadUploaded byPham van Hoan
- 9783642378867-c2Uploaded byGiriMaha
- 157 Paper MachadoUploaded byKristopher Orlowski
- Laplace's Method Approximations for Probabilistic Inference in Belief Networks With Continuous VariablesUploaded byGodofredo
- Data Collection in Steady State ConditionsUploaded byAnand Aj
- Tugas 1 (TV 10)Uploaded byVidi Fadilah
- Week 7a - Interval EstimationUploaded byLoyd Bryan Arada
- Output Case ControlUploaded byAnonymous MrXfn04
- What is a P valueUploaded byTransportasi Masal

- Estimating Electrical Relay Switching ReliabilityUploaded byarturo snodgrass
- chapter19Uploaded byapi-232613595
- 17761_work Sampling 1Uploaded byVicky Singh
- Error Propagation-Correlated VariablesUploaded byDerioUnbound
- E0EE46-BPSUBUploaded bymaheshu4u
- BayesianGuidelinesFinal(2) (1)Uploaded byZeus Herakles
- To Prepare and Validate Instrument in ResearchUploaded byAchmad Muttaqien
- MinitabUploaded bysouvik5000
- WP 2012-1 Multiple risk factors in young children's development - SABATES, R AND DEX, S.pdfUploaded byAziah Shatar
- Mathematics S P1P2 Answer Schema_PahangUploaded byKim Say Liang
- MATH1041 Final Cheat SheetUploaded byBasil Razi
- Errata - Discovering StatisticsUploaded byAa
- G 77 – 98 _RZC3.pdfUploaded bySamuel Eduardo
- CISA Lecture Domain 11Uploaded byzamanbd
- Behavioral FinanceUploaded byVALUEWALK LLC
- Sources of Validity Evidence.pdfUploaded byhbakry9608
- A Case for Health Indicators vs. Condition Indicators in Mechanical DiagnosticsUploaded byEric Bechhoefer
- 059-0058Uploaded byShaswat Singh
- 85_161_Virosil Pharma Brochure - (New)Uploaded byvk
- Non Parametric MethodUploaded byKyai Mbethik
- QUAM2012_P1.pdfUploaded byahmad Yaghi
- A Project Manager's Lessons LearnedUploaded byfredazo
- the sound of colorUploaded byjoju0531
- Data Science MCQUploaded bygaurav
- E commerce.docUploaded bybuswas
- Credit Session 4 (ESTIMATES)Uploaded byAmey Kashyap
- Clinical Trials, A Conceptual FrameworkUploaded byraluca_vior4895
- skittles term project finalUploaded byapi-254516775
- 5b Bahankuliah Ekotoksikologi Lc50lethalconcentration (1)Uploaded byWuLan Whf
- 4 Multiple ComparisonUploaded byluis d