MCMC and ABC Methodologies in the context of Controlled Branching Processes M. González, I. del Puerto Department of Mathematics. University of Extremadura Spanish Branching Processes Group Workshop Métodos Bayesianos 11 Madrid, November 2011 M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 1 / 55
108
Embed
MCMC and ABC Methodologies in the context of Controlled ...villegas/PDF/GonzalezPuerto.pdf · Contents 1 Controlled Branching Processes 2 MCMC for CBP with Deterministic Control Function
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
MCMC and ABC Methodologies in the context ofControlled Branching Processes
M. González, I. del Puerto
Department of Mathematics. University of ExtremaduraSpanish Branching Processes Group
Workshop Métodos Bayesianos 11Madrid, November 2011
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 1 / 55
Contents
1 Controlled Branching Processes
2 MCMC for CBP with Deterministic Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
3 MCMC for CBP with Random Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
4 ABC for CBP with Deterministic and Random Control FunctionsApproximate Bayesian ComputationSimulated Example
5 Concluding Remarks and ReferencesConcluding RemarksReferences
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 2 / 55
Contents
1 Controlled Branching Processes
2 MCMC for CBP with Deterministic Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
3 MCMC for CBP with Random Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
4 ABC for CBP with Deterministic and Random Control FunctionsApproximate Bayesian ComputationSimulated Example
5 Concluding Remarks and ReferencesConcluding RemarksReferences
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 2 / 55
Contents
1 Controlled Branching Processes
2 MCMC for CBP with Deterministic Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
3 MCMC for CBP with Random Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
4 ABC for CBP with Deterministic and Random Control FunctionsApproximate Bayesian ComputationSimulated Example
5 Concluding Remarks and ReferencesConcluding RemarksReferences
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 2 / 55
Contents
1 Controlled Branching Processes
2 MCMC for CBP with Deterministic Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
3 MCMC for CBP with Random Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
4 ABC for CBP with Deterministic and Random Control FunctionsApproximate Bayesian ComputationSimulated Example
5 Concluding Remarks and ReferencesConcluding RemarksReferences
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 2 / 55
Contents
1 Controlled Branching Processes
2 MCMC for CBP with Deterministic Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
3 MCMC for CBP with Random Control FunctionBayesian Inference for Controlled Branching ProcessesA Simulation-Based Method using Gibbs Sampler
4 ABC for CBP with Deterministic and Random Control FunctionsApproximate Bayesian ComputationSimulated Example
5 Concluding Remarks and ReferencesConcluding RemarksReferences
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 2 / 55
Branching Processes
Inside the general context concerning Stochastic Models, BranchingProcesses Theory provides appropriate mathematical models for descriptionof the probabilistic evolution of systems whose components (cell, particles,individuals in general), after certain life period, reproduce and die. Therefore,it can be applied in several fields (Biology, Demography, Ecology,Epidemiology, Genetics, Algorithms,...).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 3 / 55
Branching Processes
ExampleZ0 = 1
Z1 = 2Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1
Z1 = 2Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2
Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2
Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2Z2 = 7
Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2Z2 = 7
Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
ExampleZ0 = 1Z1 = 2Z2 = 7Z3 = 10
...
Zn+1 =Zn∑
j=1
Xnj
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 4 / 55
Branching Processes
Main Results for Galton–Watson Branching Processes
Let m = E[X01] and σ2 = Var[X01]
Extinction Problem
If m ≤ 1⇒ the process dies out with probability 1
If m > 1⇒ there exists a positive probability of non-extinction
Asymptotic behaviour
Statistical Inference
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 5 / 55
Branching Processes
Main Results for Galton–Watson Branching Processes
Let m = E[X01] and σ2 = Var[X01]
Extinction Problem
If m ≤ 1⇒ the process dies out with probability 1
If m > 1⇒ there exists a positive probability of non-extinction
Asymptotic behaviour
Statistical Inference
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 5 / 55
Branching Processes
Main Results for Galton–Watson Branching Processes
Let m = E[X01] and σ2 = Var[X01]
Extinction Problem
If m ≤ 1⇒ the process dies out with probability 1
If m > 1⇒ there exists a positive probability of non-extinction
Asymptotic behaviour
Statistical Inference
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 5 / 55
Branching Processes
Main Results for Galton–Watson Branching Processes
Let m = E[X01] and σ2 = Var[X01]
Extinction Problem
If m ≤ 1⇒ the process dies out with probability 1
If m > 1⇒ there exists a positive probability of non-extinction
Asymptotic behaviour
Statistical Inference
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 5 / 55
Branching Processes
Many monographs about the theory and applications about the branching processes have beenpublished:
Harris, T. (1963). The Theory of branching processes. Springer-Verlag.
Jagers, P. (1975). Branching processes with Biological Applications, John Wiley andSons, Inc.
Asmussen, S. and Hering, H. (1983). Branching processes. Birkhäuser. Boston.
Athreya, K.B. and Jagers, P. (1997). Classical and modern branching processes.Springer-Verlag.
Kimmel, M. and Axelrod, D.E. (2002). Branching processes in Biology, Springer-VerlagNew York, Inc.
Haccou, P., Jagers, P., and Vatutin, V. (2005). Branching Processes: Variation, Growth,and Extinction of Populations. Cambridge University Press.
González, M., del Puerto, I., Martínez, R., Molina, M., Mota, M., Ramos, A. (Editors)(2010). Workshop on Branching Processes and their Applications. Lecture Notes inStatistics, 197. Springer.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 6 / 55
Branching Processes
A Controlled Branching Process is a discrete-time stochastic growthpopulation model in which the individuals with reproductive capacity in eachgeneration are controlled by some function φ. This branching model iswell-suited for describing the probabilistic evolution of populations in which,for various reasons of an environmental, social or other nature, there is amechanism that establishes the number of progenitors who take part in eachgeneration.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 7 / 55
Branching Processes
Mathematically: Controlled Branching Process {Zn}n≥0
Z0 = N, Zn+1 =φn(Zn)∑
i=1
Xni, n = 0, 1, . . .
Two independent sequences of random variables (r.v.):
{Xni : i = 1, 2, . . . , n = 0, 1, . . .} are i.i.d. r.v.p = {pk : k = 0, 1, . . .} Offspring Distributionm = E[X01], σ2 = Var[X01]{φn(k) : n = 0, 1, . . . ; k = 0, 1, . . .}, where {φn(k)}k≥0 are independentstochastic processes with identical one-dimensional probabilitydistributions, n = 0, 1, . . . Random Control Functionsε(k) = E[φn(k)], σ2(k) = Var[φn(k)].φn(k) = φ(k), k = 0, 1, . . . Deterministic Control Function
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 8 / 55
Branching Processes
Mathematically: Controlled Branching Process {Zn}n≥0
Z0 = N, Zn+1 =φn(Zn)∑
i=1
Xni, n = 0, 1, . . .
Two independent sequences of random variables (r.v.):
{Xni : i = 1, 2, . . . , n = 0, 1, . . .} are i.i.d. r.v.p = {pk : k = 0, 1, . . .} Offspring Distributionm = E[X01], σ2 = Var[X01]{φn(k) : n = 0, 1, . . . ; k = 0, 1, . . .}, where {φn(k)}k≥0 are independentstochastic processes with identical one-dimensional probabilitydistributions, n = 0, 1, . . . Random Control Functionsε(k) = E[φn(k)], σ2(k) = Var[φn(k)].φn(k) = φ(k), k = 0, 1, . . . Deterministic Control Function
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 8 / 55
Sevast’yanov and Zubkov (1974)Zubkov (1974)Molina, González and Mota (1998)
Asymptotic Behaviour: Growth ratesBagley (1986)Molina, González and Mota (1998)González, Molina, del Puerto (2002, 2003, 2004, 2005a,b)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 9 / 55
Controlled Branching Processes
Main Topics InvestigatedStatistical Inference
Dion, J. P. and Essebbar, B. (1995). On the statistics of controlled branching processes. LectureNotes in Statistics, 99:14-21.
M. González, R. Martínez, I. Del Puerto (2004). Nonparametric estimation of the offspringdistribution and the mean for a controlled branching process. Test, 13(2), 465-479.
M. González, R. Martínez, I. Del Puerto (2005). Estimation of the variance for a controlledbranching process. Test, 14(1), 199-213.
T.N. Sriram, A. Bhattacharya, M. González, R. Martínez, I. Del Puerto (2007). Estimation of theoffspring mean in a controlled branching process with a random control function. StochasticProcesses and their Applications, 117, 928-946.
R. Martínez, I. del Puerto, M. Mota (2009). On asymptotic posterior normality for controlledbranching processes. Statistics, 43, 367-378.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 10 / 55
Bayesian Inference for Controlled Branching Processes
Non-Parametric FrameworkOffspring Distribution: p = {pk : k ∈ S} S finite.Deterministic Control Function: φ(·)Sample: The entire family tree up to the current generation
{Xki : i = 1, . . . , φ(Zk), k = 0, 1, . . . , n}
or at leastZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
where Zj(k) =∑φ(Zj)
i=1 I{Xji=k} = number of parents in the jth-generationwhich generate exactly k offspringObjective: Make inference on p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 11 / 55
Bayesian Inference for Controlled Branching Processes
Non-Parametric FrameworkOffspring Distribution: p = {pk : k ∈ S} S finite.Deterministic Control Function: φ(·)Sample: The entire family tree up to the current generation
{Xki : i = 1, . . . , φ(Zk), k = 0, 1, . . . , n}
or at leastZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
where Zj(k) =∑φ(Zj)
i=1 I{Xji=k} = number of parents in the jth-generationwhich generate exactly k offspringObjective: Make inference on p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 11 / 55
Bayesian Inference for Controlled Branching Processes
Non-Parametric FrameworkOffspring Distribution: p = {pk : k ∈ S} S finite.Deterministic Control Function: φ(·)Sample: The entire family tree up to the current generation
{Xki : i = 1, . . . , φ(Zk), k = 0, 1, . . . , n}
or at leastZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
where Zj(k) =∑φ(Zj)
i=1 I{Xji=k} = number of parents in the jth-generationwhich generate exactly k offspringObjective: Make inference on p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 11 / 55
Bayesian Inference for Controlled Branching Processes
Non-Parametric FrameworkOffspring Distribution: p = {pk : k ∈ S} S finite.Deterministic Control Function: φ(·)Sample: The entire family tree up to the current generation
{Xki : i = 1, . . . , φ(Zk), k = 0, 1, . . . , n}
or at leastZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
where Zj(k) =∑φ(Zj)
i=1 I{Xji=k} = number of parents in the jth-generationwhich generate exactly k offspringObjective: Make inference on p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 11 / 55
Bayesian Inference for Controlled Branching Processes
Likelihood Function
f (Zn|p) ∝∏k∈S
p∑n
j=0 Zj(k)
k
Conjugate Class of Distributions: Dirichlet FamilyPrior Distribution: p ∼ D(αk : k ∈ S)Posterior Distribution:
p|Zn ∼ D(αk +n∑
j=0
Zj(k) : k ∈ S)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 12 / 55
Bayesian Inference for Controlled Branching Processes
Likelihood Function
f (Zn|p) ∝∏k∈S
p∑n
j=0 Zj(k)
k
Conjugate Class of Distributions: Dirichlet FamilyPrior Distribution: p ∼ D(αk : k ∈ S)Posterior Distribution:
p|Zn ∼ D(αk +n∑
j=0
Zj(k) : k ∈ S)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 12 / 55
Bayesian Inference for Controlled Branching Processes
Setting out the ProblemIn real problems it is difficult to observe the entire family tree{Xki : i = 1, 2, . . . , k = 0, 1, . . . , n} or even the random variablesZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
SolutionWe introduce an algorithm to approximate the distribution
p|Z∗n
using Markov Chain Monte Carlo Methods
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 13 / 55
Bayesian Inference for Controlled Branching Processes
Setting out the ProblemIn real problems it is difficult to observe the entire family tree{Xki : i = 1, 2, . . . , k = 0, 1, . . . , n} or even the random variablesZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
SolutionWe introduce an algorithm to approximate the distribution
p|Z∗n
using Markov Chain Monte Carlo Methods
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 13 / 55
Bayesian Inference for Controlled Branching Processes
Setting out the ProblemIn real problems it is difficult to observe the entire family tree{Xki : i = 1, 2, . . . , k = 0, 1, . . . , n} or even the random variablesZn = {Zj(k) : k ∈ S, j = 0, . . . , n}
SolutionWe introduce an algorithm to approximate the distribution
p|Z∗n
using Markov Chain Monte Carlo Methods
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 13 / 55
Gibbs Sampler: Introducing the Method
Sample: Z∗n = {Zj : j = 0, . . . , n}
The Problem
p|Z∗n
Latent Variables:
Zn = {Zj(k) : k ∈ S, j = 0, . . . , n}
Gibbs Sampler:p|Zn,Z∗n Zn|Z∗n , p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 14 / 55
Gibbs Sampler: Introducing the Method
Sample: Z∗n = {Zj : j = 0, . . . , n}
The Problem
p|Z∗n
Latent Variables:
Zn = {Zj(k) : k ∈ S, j = 0, . . . , n}
Gibbs Sampler:p|Zn,Z∗n Zn|Z∗n , p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 14 / 55
Gibbs Sampler: Introducing the Method
Sample: Z∗n = {Zj : j = 0, . . . , n}
The Problem
p|Z∗n
Latent Variables:
Zn = {Zj(k) : k ∈ S, j = 0, . . . , n}
Gibbs Sampler:p|Zn,Z∗n Zn|Z∗n , p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 14 / 55
Gibbs Sampler: Introducing the Method
Sample: Z∗n = {Zj : j = 0, . . . , n}
The Problem
p|Z∗n
Latent Variables:
Zn = {Zj(k) : k ∈ S, j = 0, . . . , n}
Gibbs Sampler:p|Zn,Z∗n Zn|Z∗n , p
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 14 / 55
Gibbs Sampler: Introducing the Method
First Conditional Distribution: p|Zn,Z∗n
p|Zn,Z∗n ≡ p|Zn ∼ D(αk +n∑
j=0
Zj(k) : k ∈ S)
For j = 0, . . . , nφ(Zj) =
∑k∈S
Zj(k)
Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 15 / 55
Gibbs Sampler: Introducing the Method
First Conditional Distribution: p|Zn,Z∗n
p|Zn,Z∗n ≡ p|Zn ∼ D(αk +n∑
j=0
Zj(k) : k ∈ S)
For j = 0, . . . , nφ(Zj) =
∑k∈S
Zj(k)
Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 15 / 55
Gibbs Sampler: Introducing the Method
First Conditional Distribution: p|Zn,Z∗n
p|Zn,Z∗n ≡ p|Zn ∼ D(αk +n∑
j=0
Zj(k) : k ∈ S)
For j = 0, . . . , nφ(Zj) =
∑k∈S
Zj(k)
Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 15 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p
f (Zn|Z∗n , p) =n∏
j=0
f (Zj(k) : k ∈ S|Zj,Zj+1, p)
(Zj(k) : k ∈ S)|Zj,Zj+1, p
is obtained from aMultinomial(φ(Zj), p)
normalized by considering the constraint
Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 16 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p
f (Zn|Z∗n , p) =n∏
j=0
f (Zj(k) : k ∈ S|Zj,Zj+1, p)
(Zj(k) : k ∈ S)|Zj,Zj+1, p
is obtained from aMultinomial(φ(Zj), p)
normalized by considering the constraint
Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 16 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , pp
φ(Z0)Z0(k), k ∈ S
Z1 φ(Z1)Z1(k), k ∈ S
Z2 φ(Z2)...
......
Zn φ(Zn)Zn(k), k ∈ S
Zn+1
φ(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 17 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , pp
φ(Z0)Z0(k), k ∈ S
Z1 φ(Z1)Z1(k), k ∈ S
Z2 φ(Z2)...
......
Zn φ(Zn)Zn(k), k ∈ S
Zn+1
φ(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 17 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , pp
φ(Z0)Z0(k), k ∈ S
Z1 φ(Z1)Z1(k), k ∈ S
Z2 φ(Z2)...
......
Zn φ(Zn)Zn(k), k ∈ S
Zn+1
φ(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 17 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , pp
φ(Z0)Z0(k), k ∈ S
Z1 φ(Z1)Z1(k), k ∈ S
Z2 φ(Z2)...
......
Zn φ(Zn)Zn(k), k ∈ S
Zn+1
φ(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 17 / 55
Gibbs Sampler: Developing the Method
Algorithm
Fixed p(0)
Do l = 1Generate Z(l)
n ∼ Zn|Z∗n , p(l−1)
Generate p(l) ∼ p|Z(l)n
Do l = l + 1
For a run of the sequence {p(l)}l≥0, we choose Q + 1 vectors in the way{p(N), p(N+G), . . . , p(N+QG)}, where N is the burn-in period and G is a batchsize.
The vectors {p(N), p(N+G), . . . , p(N+QG)} are considered independent samplesfrom p|Z∗n if G and N are large enough (Tierney (1994)).
Since these vectors could be affected by the initial state p(0), we apply thealgorithm T times, obtaining a final sample of length T(Q + 1).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 18 / 55
Gibbs Sampler: Developing the Method
Algorithm
Fixed p(0)
Do l = 1Generate Z(l)
n ∼ Zn|Z∗n , p(l−1)
Generate p(l) ∼ p|Z(l)n
Do l = l + 1
For a run of the sequence {p(l)}l≥0, we choose Q + 1 vectors in the way{p(N), p(N+G), . . . , p(N+QG)}, where N is the burn-in period and G is a batchsize.
The vectors {p(N), p(N+G), . . . , p(N+QG)} are considered independent samplesfrom p|Z∗n if G and N are large enough (Tierney (1994)).
Since these vectors could be affected by the initial state p(0), we apply thealgorithm T times, obtaining a final sample of length T(Q + 1).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 18 / 55
Gibbs Sampler: Developing the Method
Algorithm
Fixed p(0)
Do l = 1Generate Z(l)
n ∼ Zn|Z∗n , p(l−1)
Generate p(l) ∼ p|Z(l)n
Do l = l + 1
For a run of the sequence {p(l)}l≥0, we choose Q + 1 vectors in the way{p(N), p(N+G), . . . , p(N+QG)}, where N is the burn-in period and G is a batchsize.
The vectors {p(N), p(N+G), . . . , p(N+QG)} are considered independent samplesfrom p|Z∗n if G and N are large enough (Tierney (1994)).
Since these vectors could be affected by the initial state p(0), we apply thealgorithm T times, obtaining a final sample of length T(Q + 1).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 18 / 55
Gibbs Sampler: Developing the Method
Algorithm
Fixed p(0)
Do l = 1Generate Z(l)
n ∼ Zn|Z∗n , p(l−1)
Generate p(l) ∼ p|Z(l)n
Do l = l + 1
For a run of the sequence {p(l)}l≥0, we choose Q + 1 vectors in the way{p(N), p(N+G), . . . , p(N+QG)}, where N is the burn-in period and G is a batchsize.
The vectors {p(N), p(N+G), . . . , p(N+QG)} are considered independent samplesfrom p|Z∗n if G and N are large enough (Tierney (1994)).
Since these vectors could be affected by the initial state p(0), we apply thealgorithm T times, obtaining a final sample of length T(Q + 1).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 18 / 55
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 30 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Introducing the Method
Second Conditional Distribution: Zn|Z∗n , p, θ(p, θ)
Z0 φ0(Z0)Z0(k), k ∈ S
Z1 φ1(Z1)Z1(k), k ∈ S
Z2 φ2(Z2)...
......
Zn φn(Zn)Zn(k), k ∈ S
Zn+1
φj(Zj) =∑k∈S
Zj(k), Zj+1 =∑k∈S
kZj(k)
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 31 / 55
Gibbs Sampler: Developing the Method
AlgorithmInitialize l = 0Generate p(0) ∼ Dirichlet(α)Generate θ(0) from π(θ) = ϕ(a, b)−1θa/Ab(θ)Iterate
l = l + 1Generate Z(l)
n ∼ f (Zn | Z∗n , p(l−1), θ(l−1))Generate (p(l), θ(l)) ∼ π(p, θ | Z(l)
n )
For a run of the sequence {(θ, p)(l)}l≥0, we choose Q + 1 vectors in the way{(θ, p)(N), (θ, p)(N+G)), . . . , (θ, p)(N+QG)}, where N is a burning period and Gis a batch size.
The vectors {(θ, p)(N), (θ, p)(N+G), . . . , (θ, p)(N+QG)} are consideredindependent samples from (θ, p)|Z∗n if G and N are large enough.
Since these vectors could be affected by the initial state (θ, p)(0), we apply thealgorithm T times, obtaining a final sample of length T(Q + 1).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 32 / 55
Gibbs Sampler: Developing the Method
AlgorithmInitialize l = 0Generate p(0) ∼ Dirichlet(α)Generate θ(0) from π(θ) = ϕ(a, b)−1θa/Ab(θ)Iterate
l = l + 1Generate Z(l)
n ∼ f (Zn | Z∗n , p(l−1), θ(l−1))Generate (p(l), θ(l)) ∼ π(p, θ | Z(l)
n )
For a run of the sequence {(θ, p)(l)}l≥0, we choose Q + 1 vectors in the way{(θ, p)(N), (θ, p)(N+G)), . . . , (θ, p)(N+QG)}, where N is a burning period and Gis a batch size.
The vectors {(θ, p)(N), (θ, p)(N+G), . . . , (θ, p)(N+QG)} are consideredindependent samples from (θ, p)|Z∗n if G and N are large enough.
Since these vectors could be affected by the initial state (θ, p)(0), we apply thealgorithm T times, obtaining a final sample of length T(Q + 1).
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 32 / 55
Control function: φ(x) = 7 if x ≤ 7; x if 7 < x ≤ 20; 20 if x > 20
Observed Data: n = 40
0 10 20 30 40
510
1520
25
Generations
Indi
vidu
als
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 40 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 10
●
●
●
●
●
●
●
●
●
●
●
0.10 0.12 0.14 0.16 0.18 0.20
5000
1000
015
000
2000
025
000
3000
0
Tolerance
effe
ctiv
e sa
mpl
e si
ze o
ver
5 m
illio
ns o
f sim
ulat
ions
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 41 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 10
0.8 1.0 1.2 1.4
01
23
45
Tolerance: 0.11
Den
sity
0.6 0.8 1.0 1.2 1.4
01
23
45
Tolerance: 0.12
Den
sity
0.6 0.8 1.0 1.2 1.40
12
34
5
Tolerance: 0.13
Den
sity
0.6 0.8 1.0 1.2 1.4
01
23
45
Tolerance: 0.14
Den
sity
0.6 0.8 1.0 1.2 1.4
01
23
45
Tolerance: 0.15
Den
sity
0.6 0.8 1.0 1.2 1.4 1.6
01
23
45
Tolerance: 0.16
Den
sity
0.6 0.8 1.0 1.2 1.4 1.6
01
23
45
Tolerance: 0.17
Den
sity
0.6 0.8 1.0 1.2 1.4 1.6
01
23
45
Tolerance: 0.18
Den
sity
0.6 0.8 1.0 1.2 1.4 1.6
01
23
45
Tolerance: 0.19
Den
sity
0.6 0.8 1.0 1.2 1.4 1.6
01
23
45
Tolerance: 0.2
Den
sity
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 42 / 55
ABC: Simulated Example
Sample Information: Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 10
● ●
●
●
●
●
●
●
●
●
0.12 0.14 0.16 0.18 0.20
0.05
0.06
0.07
0.08
0.09
0.10
0.11
Tolerance
ISE
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 43 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 20
●
●
●
●
●
●
●
●
●
●
●
0.10 0.12 0.14 0.16 0.18 0.20
2000
040
000
6000
080
000
Tolerance
effe
ctiv
e sa
mpl
e si
ze o
ver
10 m
illio
ns o
f sim
ulat
ions
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 44 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 20
0.9 1.0 1.1 1.2 1.3 1.4
02
46
8
Tolerance: 0.11
Den
sity
0.9 1.0 1.1 1.2 1.3 1.4
02
46
8
Tolerance: 0.12
Den
sity
0.9 1.0 1.1 1.2 1.3 1.40
24
68
Tolerance: 0.13
Den
sity
0.9 1.0 1.1 1.2 1.3 1.4 1.5
02
46
8
Tolerance: 0.14
Den
sity
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
02
46
8
Tolerance: 0.15
Den
sity
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
02
46
8
Tolerance: 0.16
Den
sity
0.8 1.0 1.2 1.4
02
46
8
Tolerance: 0.17
Den
sity
0.8 1.0 1.2 1.4
02
46
8
Tolerance: 0.18
Den
sity
0.8 1.0 1.2 1.4
02
46
8
Tolerance: 0.19
Den
sity
0.8 1.0 1.2 1.4
02
46
8
Tolerance: 0.2
Den
sity
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 45 / 55
ABC: Simulated Example
Sample Information: Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 20
●
●●
●
●
●
●
●
●
●
0.12 0.14 0.16 0.18 0.20
0.25
0.30
0.35
0.40
0.45
0.50
Tolerance
ISE
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 46 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 30
● ● ●●
●
●
●
●
●
●
●
0.10 0.12 0.14 0.16 0.18 0.20
010
000
2000
030
000
4000
050
000
6000
0
Tolerance
effe
ctiv
e sa
mpl
e si
ze o
ver
20 m
illio
ns o
f sim
ulat
ions
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 47 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 30
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.11
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.12
Den
sity
0.9 1.0 1.1 1.20
24
68
Tolerance: 0.13
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.14
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.15
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.16
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.17
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.18
Den
sity
0.8 0.9 1.0 1.1 1.2 1.3 1.4
02
46
8
Tolerance: 0.19
Den
sity
0.8 0.9 1.0 1.1 1.2 1.3 1.4
02
46
8
Tolerance: 0.2
Den
sity
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 48 / 55
ABC: Simulated Example
Sample Information: Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 30
●
●
●
●
●
●
●
●
●
●
0.12 0.14 0.16 0.18 0.20
0.05
0.10
0.15
0.20
0.25
0.30
Tolerance
ISE
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 49 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 40
● ● ● ●●
●
●
●
●
●
●
0.10 0.12 0.14 0.16 0.18 0.20
050
0010
000
1500
0
Tolerance
effe
ctiv
e sa
mpl
e si
ze o
ver
20 m
illio
ns o
f sim
ulat
ions
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 50 / 55
ABC: Simulated Example
Sample Information Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 40
0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20
02
46
8
Tolerance: 0.13
Den
sity
0.9 1.0 1.1 1.2
02
46
8
Tolerance: 0.14
Den
sity
0.9 1.0 1.1 1.2
02
46
8
Tolerance: 0.15
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.16
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.17
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.18
Den
sity
0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.19
Den
sity
0.8 0.9 1.0 1.1 1.2 1.3
02
46
8
Tolerance: 0.2
Den
sity
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 51 / 55
ABC: Simulated Example
Sample Information: Z∗n , p ∼ D(1/2, . . . , 1/2), N = 20 millionsGeneration 40
●
●
●
●
●
●
●
●
0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.20
0.26
0.28
0.30
0.32
0.34
0.36
0.38
Tolerance
ISE
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 52 / 55
Concluding Remarks
In a non-parametric Bayesian framework we can make inference on theoffspring distribution of CBP, and consequently on the rest of offspringparameters, without observing the entire family tree, but onlyconsidering the total number of individuals in each generation.
We use a MCMC method (Gibbs sampler) in order to give a "likely"approach to family trees, for both CBP with deterministic and withrandom control function.
We take advantage of the ABC methodology to make inference on themain parameters of the model by simulating.
The ABC approach shows a quite good behaviour, being a goodalternative to the MCMC approach.
We have developed the above methodologies using statistical softwareand programming environment R.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 53 / 55
Concluding Remarks
In a non-parametric Bayesian framework we can make inference on theoffspring distribution of CBP, and consequently on the rest of offspringparameters, without observing the entire family tree, but onlyconsidering the total number of individuals in each generation.
We use a MCMC method (Gibbs sampler) in order to give a "likely"approach to family trees, for both CBP with deterministic and withrandom control function.
We take advantage of the ABC methodology to make inference on themain parameters of the model by simulating.
The ABC approach shows a quite good behaviour, being a goodalternative to the MCMC approach.
We have developed the above methodologies using statistical softwareand programming environment R.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 53 / 55
Concluding Remarks
In a non-parametric Bayesian framework we can make inference on theoffspring distribution of CBP, and consequently on the rest of offspringparameters, without observing the entire family tree, but onlyconsidering the total number of individuals in each generation.
We use a MCMC method (Gibbs sampler) in order to give a "likely"approach to family trees, for both CBP with deterministic and withrandom control function.
We take advantage of the ABC methodology to make inference on themain parameters of the model by simulating.
The ABC approach shows a quite good behaviour, being a goodalternative to the MCMC approach.
We have developed the above methodologies using statistical softwareand programming environment R.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 53 / 55
Concluding Remarks
In a non-parametric Bayesian framework we can make inference on theoffspring distribution of CBP, and consequently on the rest of offspringparameters, without observing the entire family tree, but onlyconsidering the total number of individuals in each generation.
We use a MCMC method (Gibbs sampler) in order to give a "likely"approach to family trees, for both CBP with deterministic and withrandom control function.
We take advantage of the ABC methodology to make inference on themain parameters of the model by simulating.
The ABC approach shows a quite good behaviour, being a goodalternative to the MCMC approach.
We have developed the above methodologies using statistical softwareand programming environment R.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 53 / 55
Concluding Remarks
In a non-parametric Bayesian framework we can make inference on theoffspring distribution of CBP, and consequently on the rest of offspringparameters, without observing the entire family tree, but onlyconsidering the total number of individuals in each generation.
We use a MCMC method (Gibbs sampler) in order to give a "likely"approach to family trees, for both CBP with deterministic and withrandom control function.
We take advantage of the ABC methodology to make inference on themain parameters of the model by simulating.
The ABC approach shows a quite good behaviour, being a goodalternative to the MCMC approach.
We have developed the above methodologies using statistical softwareand programming environment R.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 53 / 55
References
Bagley, J. H. (1986). On the almost sure convergence of controlled branching processes. Journal of Applied Probability, 23:827-831.
M. González, M. Molina, I. del Puerto (2002). On the class of controlled branching process with random control functions. Journal ofApplied Probability, 39 (4), 804-815.
M. González, M. Molina, I. del Puerto (2003). On the geometric growth in controlled branching processes with random controlfunction. Journal of Applied Probability, 40(4), 995-1006.
M. González, M. Molina, I. del Puerto (2004). Limiting distribution for subcritical controlled branching processes with randomcontrol function. Statistics and Probability Letters, 67(3), 277-284.
M. González, M. Molina, I. del Puerto (2005a). Asymptotic behaviour of critical controlled branching process with random controlfunction. Journal of Applied Probability, 42(2), 463-477.
M. González, M. Molina, I. del Puerto (2005b). On the L2-convergence of controlled branching processes with random controlfunction. Bernoulli, 11(1), 37-46.
M. Molina, M. González, M. Mota (1998). Some theoretical results about superadditive controlled Galton-Watson branchingprocesses. Proceedings of the International Conference Prague Stochastics98, 2, 413-418.
Sevast’yanov, B. A. and Zubkov, A. (1974). Controlled branching processes. Theory of Probability and its Applications, 19:14-24.
Tierney, L. (1994). Markov chains for exploring posterior distributions. Annals of Statistics, 22, 1701-1762.
Wilkinson, R.D. (2008). Approximate Bayesian computation (ABC) gives exact results under the assumption of model error.Technical Report. arXiv:0811.3355.
Zubkov, A. M. (1974). Analogies between Galton-Watson processes and φ-branching processes. Theory of Probability and itsApplications,19:309-331.
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 54 / 55
Thank you very much!
M. González (Universidad de Extremadura) Bayesian Inference for CBP Workshop Métodos Bayesianos 11 55 / 55