Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

ABC and empirical likelihood

2,088 views

Published on

talk at SuSTain Structure and uncertainty
modelling, inference and computation in complex stochastic systems, Bristol, 24-27 Sept., 2012

  • Be the first to comment

ABC and empirical likelihood

  1. 1. Approximate Bayesian Computation (ABC) and empirical likelihood Christian P. Robert Structure and uncertainty, Bristol, Sept. 25, 2012 Universit´ Paris-Dauphine, IuF, & CREST e Joint work with Kerrie L. Mengersen and P. Pudlo
  2. 2. OutlineIntroductionABCABC as an inference machineABCel
  3. 3. Intractable likelihood Case of a well-defined statistical model where the likelihood function (θ|y) = f (y1 , . . . , yn |θ) is (really!) not available in closed form can (easily!) be neither completed nor demarginalised cannot be estimated by an unbiased estimator c Prohibits direct implementation of a generic MCMC algorithm like Metropolis–Hastings
  4. 4. Intractable likelihood Case of a well-defined statistical model where the likelihood function (θ|y) = f (y1 , . . . , yn |θ) is (really!) not available in closed form can (easily!) be neither completed nor demarginalised cannot be estimated by an unbiased estimator c Prohibits direct implementation of a generic MCMC algorithm like Metropolis–Hastings
  5. 5. Different perspectives on abc What is the (most) fundamental issue? a mere computational issue (that will eventually end up being solved by more powerful computers, &tc, even if too costly in the short term) an inferential issue (opening opportunities for new inference machine, with different legitimity than classical B approach) a Bayesian conundrum (while inferencial methods available, how closely related to the B approach?)
  6. 6. Different perspectives on abc What is the (most) fundamental issue? a mere computational issue (that will eventually end up being solved by more powerful computers, &tc, even if too costly in the short term) an inferential issue (opening opportunities for new inference machine, with different legitimity than classical B approach) a Bayesian conundrum (while inferencial methods available, how closely related to the B approach?)
  7. 7. Different perspectives on abc What is the (most) fundamental issue? a mere computational issue (that will eventually end up being solved by more powerful computers, &tc, even if too costly in the short term) an inferential issue (opening opportunities for new inference machine, with different legitimity than classical B approach) a Bayesian conundrum (while inferencial methods available, how closely related to the B approach?)
  8. 8. Econom’ections Similar exploration of simulation-based and approximation techniques in Econometrics Simulated method of moments Method of simulated moments Simulated pseudo-maximum-likelihood Indirect inference [Gouri´roux & Monfort, 1996] e even though motivation is partly-defined models rather than complex likelihoods
  9. 9. Econom’ections Similar exploration of simulation-based and approximation techniques in Econometrics Simulated method of moments Method of simulated moments Simulated pseudo-maximum-likelihood Indirect inference [Gouri´roux & Monfort, 1996] e even though motivation is partly-defined models rather than complex likelihoods
  10. 10. Indirect inference ^ Minimise [in θ] a distance between estimators β based on a pseudo-model for genuine observations and for observations simulated under the true model and the parameter θ. [Gouri´roux, Monfort, & Renault, 1993; e Smith, 1993; Gallant & Tauchen, 1996]
  11. 11. Indirect inference (PML vs. PSE) Example of the pseudo-maximum-likelihood (PML) ^ β(y) = arg max log f (yt |β, y1:(t−1) ) β t leading to arg min ||β(yo ) − β(y1 (θ), . . . , yS (θ))||2 ^ ^ θ when ys (θ) ∼ f (y|θ) s = 1, . . . , S
  12. 12. Indirect inference (PML vs. PSE) Example of the pseudo-score-estimator (PSE) 2 ∂ log f ^ β(y) = arg min (yt |β, y1:(t−1) ) β t ∂β leading to arg min ||β(yo ) − β(y1 (θ), . . . , yS (θ))||2 ^ ^ θ when ys (θ) ∼ f (y|θ) s = 1, . . . , S
  13. 13. Consistent indirect inference “...in order to get a unique solution the dimension of the auxiliary parameter β must be larger than or equal to the dimension of the initial parameter θ. If the problem is just identified the different methods become easier...” Consistency depending on the criterion and on the asymptotic identifiability of θ [Gouri´roux & Monfort, 1996, p. 66] e
  14. 14. Consistent indirect inference “...in order to get a unique solution the dimension of the auxiliary parameter β must be larger than or equal to the dimension of the initial parameter θ. If the problem is just identified the different methods become easier...” Consistency depending on the criterion and on the asymptotic identifiability of θ [Gouri´roux & Monfort, 1996, p. 66] e
  15. 15. Choice of pseudo-model Arbitrariness of pseudo-model Pick model such that ^ 1. β(θ) not flat (i.e. sensitive to changes in θ) ^ 2. β(θ) not dispersed (i.e. robust agains changes in ys (θ)) [Frigessi & Heggland, 2004]
  16. 16. Approximate Bayesian computation Introduction ABC Genesis of ABC ABC basics Advances and interpretations ABC as knn ABC as an inference machine ABCel
  17. 17. Genetic background of ABC skip genetics ABC is a recent computational technique that only requires being able to sample from the likelihood f (·|θ) This technique stemmed from population genetics models, about 15 years ago, and population geneticists still contribute significantly to methodological developments of ABC. [Griffith & al., 1997; Tavar´ & al., 1999] e
  18. 18. Demo-genetic inference Each model is characterized by a set of parameters θ that cover historical (time divergence, admixture time ...), demographics (population sizes, admixture rates, migration rates, ...) and genetic (mutation rate, ...) factors The goal is to estimate these parameters from a dataset of polymorphism (DNA sample) y observed at the present time Problem: most of the time, we cannot calculate the likelihood of the polymorphism data f (y|θ)...
  19. 19. Demo-genetic inference Each model is characterized by a set of parameters θ that cover historical (time divergence, admixture time ...), demographics (population sizes, admixture rates, migration rates, ...) and genetic (mutation rate, ...) factors The goal is to estimate these parameters from a dataset of polymorphism (DNA sample) y observed at the present time Problem: most of the time, we cannot calculate the likelihood of the polymorphism data f (y|θ)...
  20. 20. Neutral model at a given microsatellite locus, in a closedpanmictic population at equilibrium Mutations according to the Simple stepwise Mutation Model (SMM) • date of the mutations ∼ Poisson process with intensity θ/2 over the branches • MRCA = 100 • independent mutations: ±1 with pr. 1/2 Sample of 8 genes
  21. 21. Neutral model at a given microsatellite locus, in a closedpanmictic population at equilibrium Kingman’s genealogy When time axis is normalized, T (k) ∼ Exp(k(k − 1)/2) Mutations according to the Simple stepwise Mutation Model (SMM) • date of the mutations ∼ Poisson process with intensity θ/2 over the branches • MRCA = 100 • independent mutations: ±1 with pr. 1/2
  22. 22. Neutral model at a given microsatellite locus, in a closedpanmictic population at equilibrium Kingman’s genealogy When time axis is normalized, T (k) ∼ Exp(k(k − 1)/2) Mutations according to the Simple stepwise Mutation Model (SMM) • date of the mutations ∼ Poisson process with intensity θ/2 over the branches • MRCA = 100 • independent mutations: ±1 with pr. 1/2
  23. 23. Neutral model at a given microsatellite locus, in a closedpanmictic population at equilibrium Kingman’s genealogy When time axis is normalized, T (k) ∼ Exp(k(k − 1)/2) Mutations according to the Simple stepwise Mutation Model (SMM) • date of the mutations ∼ Poisson process with intensity θ/2 over the branches Observations: leafs of the tree • MRCA = 100 ^ θ=? • independent mutations: ±1 with pr. 1/2
  24. 24. Much more interesting models. . . several independent locus Independent gene genealogies and mutations different populations linked by an evolutionary scenario made of divergences, admixtures, migrations between populations, etc. larger sample size usually between 50 and 100 genes MRCA τ2 τ1 A typical evolutionary scenario: POP 0 POP 1 POP 2
  25. 25. Intractable likelihood Missing (too missing!) data structure: f (y|θ) = f (y|G , θ)f (G |θ)dG G cannot be computed in a manageable way... The genealogies are considered as nuisance parameters This modelling clearly differs from the phylogenetic perspective where the tree is the parameter of interest.
  26. 26. Intractable likelihood Missing (too missing!) data structure: f (y|θ) = f (y|G , θ)f (G |θ)dG G cannot be computed in a manageable way... The genealogies are considered as nuisance parameters This modelling clearly differs from the phylogenetic perspective where the tree is the parameter of interest.
  27. 27. not-so-obvious ancestry... You went to school to learn, girl (. . . ) Why 2 plus 2 makes four Now, now, now, I’m gonna teach you (. . . ) All you gotta do is repeat after me! A, B, C! It’s easy as 1, 2, 3! Or simple as Do, Re, Mi! (. . . )
  28. 28. A?B?C? A stands for approximate [wrong likelihood / picture] B stands for Bayesian C stands for computation [producing a parameter sample]
  29. 29. A?B?C? A stands for approximate [wrong likelihood / picture] B stands for Bayesian C stands for computation [producing a parameter sample]
  30. 30. A?B?C? ESS=108.9 ESS=81.48 ESS=105.2 3.0 2.0 2.0 Density Density Density 1.5 1.0 1.0 0.0 0.0 0.0 A stands for approximate −0.4 0.0 0.2 ESS=133.3 θ 0.4 0.6 −0.8 −0.6 −0.4 ESS=87.75 θ −0.2 0.0 −0.2 0.0 0.2 ESS=72.89 θ 0.4 0.6 0.8 3.0 2.0 4 [wrong likelihood / 2.0 Density Density Density 1.0 1.0 2 0.0 0.0 0 picture] −0.8 −0.4 ESS=116.5 θ 0.0 0.2 0.4 −0.2 0.0 0.2 ESS=103.9 θ 0.4 0.6 0.8 −0.2 0.0 ESS=126.9 θ 0.2 0.4 3.0 2.0 3.0 2.0 Density Density Density 1.0 1.5 1.0 B stands for Bayesian 0.0 0.0 0.0 −0.4 0.0 0.2 0.4 0.6 −0.4 −0.2 0.0 0.2 0.4 −0.8 −0.4 0.0 0.4 ESS=113.3 θ ESS=92.99 θ ESS=121.4 θ 3.0 2.0 C stands for computation 2.0 Density Density Density 1.5 1.0 1.0 0.0 0.0 0.0 [producing a parameter −0.6 −0.2 ESS=133.6 θ 0.2 0.6 −0.2 0.0 0.2 ESS=116.4 θ 0.4 0.6 −0.5 ESS=131.6 θ 0.0 0.5 0.0 1.0 2.0 3.0 2.0 sample] Density Density Density 1.0 1.0 0.0 0.0 −0.6 −0.2 0.2 0.4 0.6 −0.4 −0.2 0.0 0.2 0.4 −0.5 0.0 0.5
  31. 31. How Bayesian is aBc? Could we turn the resolution into a Bayesian answer? ideally so (not meaningfull: requires ∞-ly powerful computer asymptotically so (when sample size goes to ∞: meaningfull?) approximation error unknown (w/o costly simulation) true Bayes for wrong model (formal and artificial) true Bayes for estimated likelihood (back to econometrics?)
  32. 32. Untractable likelihoodBack to stage zero: what can we dowhen a likelihood function f (y|θ) iswell-defined but impossible / toocostly to compute...? MCMC cannot be implemented! shall we give up Bayesian inference altogether?! or settle for an almost Bayesian inference/picture...?
  33. 33. Untractable likelihoodBack to stage zero: what can we dowhen a likelihood function f (y|θ) iswell-defined but impossible / toocostly to compute...? MCMC cannot be implemented! shall we give up Bayesian inference altogether?! or settle for an almost Bayesian inference/picture...?
  34. 34. ABC methodology Bayesian setting: target is π(θ)f (x|θ) When likelihood f (x|θ) not in closed form, likelihood-free rejection technique: Foundation For an observation y ∼ f (y|θ), under the prior π(θ), if one keeps jointly simulating θ ∼ π(θ) , z ∼ f (z|θ ) , until the auxiliary variable z is equal to the observed value, z = y, then the selected θ ∼ π(θ|y) [Rubin, 1984; Diggle & Gratton, 1984; Tavar´ et al., 1997] e
  35. 35. ABC methodology Bayesian setting: target is π(θ)f (x|θ) When likelihood f (x|θ) not in closed form, likelihood-free rejection technique: Foundation For an observation y ∼ f (y|θ), under the prior π(θ), if one keeps jointly simulating θ ∼ π(θ) , z ∼ f (z|θ ) , until the auxiliary variable z is equal to the observed value, z = y, then the selected θ ∼ π(θ|y) [Rubin, 1984; Diggle & Gratton, 1984; Tavar´ et al., 1997] e
  36. 36. ABC methodology Bayesian setting: target is π(θ)f (x|θ) When likelihood f (x|θ) not in closed form, likelihood-free rejection technique: Foundation For an observation y ∼ f (y|θ), under the prior π(θ), if one keeps jointly simulating θ ∼ π(θ) , z ∼ f (z|θ ) , until the auxiliary variable z is equal to the observed value, z = y, then the selected θ ∼ π(θ|y) [Rubin, 1984; Diggle & Gratton, 1984; Tavar´ et al., 1997] e
  37. 37. A as A...pproximative When y is a continuous random variable, strict equality z = y is replaced with a tolerance zone ρ(y, z) where ρ is a distance Output distributed from def π(θ) Pθ {ρ(y, z) < } ∝ π(θ|ρ(y, z) < ) [Pritchard et al., 1999]
  38. 38. A as A...pproximative When y is a continuous random variable, strict equality z = y is replaced with a tolerance zone ρ(y, z) where ρ is a distance Output distributed from def π(θ) Pθ {ρ(y, z) < } ∝ π(θ|ρ(y, z) < ) [Pritchard et al., 1999]
  39. 39. ABC algorithm In most implementations, further degree of A...pproximation: Algorithm 1 Likelihood-free rejection sampler for i = 1 to N do repeat generate θ from the prior distribution π(·) generate z from the likelihood f (·|θ ) until ρ{η(z), η(y)} set θi = θ end for where η(y) defines a (not necessarily sufficient) statistic
  40. 40. Output The likelihood-free algorithm samples from the marginal in z of: π(θ)f (z|θ)IA ,y (z) π (θ, z|y) = , A ,y ×Θ π(θ)f (z|θ)dzdθ where A ,y = {z ∈ D|ρ(η(z), η(y)) < }. The idea behind ABC is that the summary statistics coupled with a small tolerance should provide a good approximation of the posterior distribution: π (θ|y) = π (θ, z|y)dz ≈ π(θ|y) . ...does it?!
  41. 41. Output The likelihood-free algorithm samples from the marginal in z of: π(θ)f (z|θ)IA ,y (z) π (θ, z|y) = , A ,y ×Θ π(θ)f (z|θ)dzdθ where A ,y = {z ∈ D|ρ(η(z), η(y)) < }. The idea behind ABC is that the summary statistics coupled with a small tolerance should provide a good approximation of the posterior distribution: π (θ|y) = π (θ, z|y)dz ≈ π(θ|y) . ...does it?!
  42. 42. Output The likelihood-free algorithm samples from the marginal in z of: π(θ)f (z|θ)IA ,y (z) π (θ, z|y) = , A ,y ×Θ π(θ)f (z|θ)dzdθ where A ,y = {z ∈ D|ρ(η(z), η(y)) < }. The idea behind ABC is that the summary statistics coupled with a small tolerance should provide a good approximation of the posterior distribution: π (θ|y) = π (θ, z|y)dz ≈ π(θ|y) . ...does it?!
  43. 43. Output The likelihood-free algorithm samples from the marginal in z of: π(θ)f (z|θ)IA ,y (z) π (θ, z|y) = , A ,y ×Θ π(θ)f (z|θ)dzdθ where A ,y = {z ∈ D|ρ(η(z), η(y)) < }. The idea behind ABC is that the summary statistics coupled with a small tolerance should provide a good approximation of the restricted posterior distribution: π (θ|y) = π (θ, z|y)dz ≈ π(θ|η(y)) . Not so good..! skip convergence details!
  44. 44. Convergence of ABC What happens when → 0? For B ⊂ Θ, we have A f (z|θ)dz f (z|θ)π(θ)dθ ,y B π(θ)dθ = dz B A ,y ×Θ π(θ)f (z|θ)dzdθ A ,y A ,y ×Θ π(θ)f (z|θ)dzdθ B f (z|θ)π(θ)dθ m(z) = dz A ,y m(z) A ,y ×Θ π(θ)f (z|θ)dzdθ m(z) = π(B|z) dz A ,y A ,y ×Θ π(θ)f (z|θ)dzdθ which indicates convergence for a continuous π(B|z).
  45. 45. Convergence of ABC What happens when → 0? For B ⊂ Θ, we have A f (z|θ)dz f (z|θ)π(θ)dθ ,y B π(θ)dθ = dz B A ,y ×Θ π(θ)f (z|θ)dzdθ A ,y A ,y ×Θ π(θ)f (z|θ)dzdθ B f (z|θ)π(θ)dθ m(z) = dz A ,y m(z) A ,y ×Θ π(θ)f (z|θ)dzdθ m(z) = π(B|z) dz A ,y A ,y ×Θ π(θ)f (z|θ)dzdθ which indicates convergence for a continuous π(B|z).
  46. 46. Convergence (do not attempt!) ...and the above does not apply to insufficient statistics: If η(y) is not a sufficient statistics, the best one can hope for is π(θ|η(y)) , not π(θ|y) If η(y) is an ancillary statistic, the whole information contained in y is lost!, the “best” one can “hope” for is π(θ|η(y)) = π(θ) Bummer!!!
  47. 47. Convergence (do not attempt!) ...and the above does not apply to insufficient statistics: If η(y) is not a sufficient statistics, the best one can hope for is π(θ|η(y)) , not π(θ|y) If η(y) is an ancillary statistic, the whole information contained in y is lost!, the “best” one can “hope” for is π(θ|η(y)) = π(θ) Bummer!!!
  48. 48. Convergence (do not attempt!) ...and the above does not apply to insufficient statistics: If η(y) is not a sufficient statistics, the best one can hope for is π(θ|η(y)) , not π(θ|y) If η(y) is an ancillary statistic, the whole information contained in y is lost!, the “best” one can “hope” for is π(θ|η(y)) = π(θ) Bummer!!!
  49. 49. Convergence (do not attempt!) ...and the above does not apply to insufficient statistics: If η(y) is not a sufficient statistics, the best one can hope for is π(θ|η(y)) , not π(θ|y) If η(y) is an ancillary statistic, the whole information contained in y is lost!, the “best” one can “hope” for is π(θ|η(y)) = π(θ) Bummer!!!
  50. 50. MA example Inference on the parameters of a MA(q) model q xt = t + ϑi t−i t−i i.i.d.w.n. i=1 bypass MA illustration Simple prior: uniform over the inverse [real and complex] roots in q Q(u) = 1 − ϑi u i i=1 under the identifiability conditions
  51. 51. MA example Inference on the parameters of a MA(q) model q xt = t + ϑi t−i t−i i.i.d.w.n. i=1 bypass MA illustration Simple prior: uniform prior over the identifiability zone in the parameter space, i.e. triangle for MA(2)
  52. 52. MA example (2) ABC algorithm thus made of 1. picking a new value (ϑ1 , ϑ2 ) in the triangle 2. generating an iid sequence ( t )−q<t T 3. producing a simulated series (xt )1 t T Distance: basic distance between the series T ρ((xt )1 t T , (xt )1 t T) = (xt − xt )2 t=1 or distance between summary statistics like the q = 2 autocorrelations T τj = xt xt−j t=j+1
  53. 53. MA example (2) ABC algorithm thus made of 1. picking a new value (ϑ1 , ϑ2 ) in the triangle 2. generating an iid sequence ( t )−q<t T 3. producing a simulated series (xt )1 t T Distance: basic distance between the series T ρ((xt )1 t T , (xt )1 t T) = (xt − xt )2 t=1 or distance between summary statistics like the q = 2 autocorrelations T τj = xt xt−j t=j+1
  54. 54. Comparison of distance impact Impact of tolerance on ABC sample against either distance ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
  55. 55. Comparison of distance impact 4 1.5 3 1.0 2 0.5 1 0.0 0 0.0 0.2 0.4 0.6 0.8 −2.0 −1.0 0.0 0.5 1.0 1.5 θ1 θ2 Impact of tolerance on ABC sample against either distance ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
  56. 56. Comparison of distance impact 4 1.5 3 1.0 2 0.5 1 0.0 0 0.0 0.2 0.4 0.6 0.8 −2.0 −1.0 0.0 0.5 1.0 1.5 θ1 θ2 Impact of tolerance on ABC sample against either distance ( = 100%, 10%, 1%, 0.1%) for an MA(2) model
  57. 57. Comments Role of distance paramount (because = 0) Scaling of components of η(y) is also determinant matters little if “small enough” representative of “curse of dimensionality” small is beautiful! the data as a whole may be paradoxically weakly informative for ABC
  58. 58. ABC (simul’) advances how approximative is ABC? ABC as knn Simulating from the prior is often poor in efficiency Either modify the proposal distribution on θ to increase the density of x’s within the vicinity of y ... [Marjoram et al, 2003; Bortot et al., 2007, Sisson et al., 2007] ...or by viewing the problem as a conditional density estimation and by developing techniques to allow for larger [Beaumont et al., 2002] .....or even by including in the inferential framework [ABCµ ] [Ratmann et al., 2009]
  59. 59. ABC (simul’) advances how approximative is ABC? ABC as knn Simulating from the prior is often poor in efficiency Either modify the proposal distribution on θ to increase the density of x’s within the vicinity of y ... [Marjoram et al, 2003; Bortot et al., 2007, Sisson et al., 2007] ...or by viewing the problem as a conditional density estimation and by developing techniques to allow for larger [Beaumont et al., 2002] .....or even by including in the inferential framework [ABCµ ] [Ratmann et al., 2009]
  60. 60. ABC (simul’) advances how approximative is ABC? ABC as knn Simulating from the prior is often poor in efficiency Either modify the proposal distribution on θ to increase the density of x’s within the vicinity of y ... [Marjoram et al, 2003; Bortot et al., 2007, Sisson et al., 2007] ...or by viewing the problem as a conditional density estimation and by developing techniques to allow for larger [Beaumont et al., 2002] .....or even by including in the inferential framework [ABCµ ] [Ratmann et al., 2009]
  61. 61. ABC (simul’) advances how approximative is ABC? ABC as knn Simulating from the prior is often poor in efficiency Either modify the proposal distribution on θ to increase the density of x’s within the vicinity of y ... [Marjoram et al, 2003; Bortot et al., 2007, Sisson et al., 2007] ...or by viewing the problem as a conditional density estimation and by developing techniques to allow for larger [Beaumont et al., 2002] .....or even by including in the inferential framework [ABCµ ] [Ratmann et al., 2009]
  62. 62. ABC-NP Better usage of [prior] simulations by adjustement: instead of throwing away θ such that ρ(η(z), η(y)) > , replace θ’s with locally regressed transforms θ∗ = θ − {η(z) − η(y)}T β ^ [Csill´ry et al., TEE, 2010] e ^ where β is obtained by [NP] weighted least square regression on (η(z) − η(y)) with weights Kδ {ρ(η(z), η(y))} [Beaumont et al., 2002, Genetics]
  63. 63. ABC-NP (regression) Also found in the subsequent literature, e.g. in Fearnhead-Prangle (2012) : weight directly simulation by Kδ {ρ(η(z(θ)), η(y))} or S 1 Kδ {ρ(η(zs (θ)), η(y))} S s=1 [consistent estimate of f (η|θ)] Curse of dimensionality: poor estimate when d = dim(η) is large...
  64. 64. ABC-NP (regression) Also found in the subsequent literature, e.g. in Fearnhead-Prangle (2012) : weight directly simulation by Kδ {ρ(η(z(θ)), η(y))} or S 1 Kδ {ρ(η(zs (θ)), η(y))} S s=1 [consistent estimate of f (η|θ)] Curse of dimensionality: poor estimate when d = dim(η) is large...
  65. 65. ABC-NP (density estimation) Use of the kernel weights Kδ {ρ(η(z(θ)), η(y))} leads to the NP estimate of the posterior expectation i θi Kδ {ρ(η(z(θi )), η(y))} i Kδ {ρ(η(z(θi )), η(y))} [Blum, JASA, 2010]
  66. 66. ABC-NP (density estimation) Use of the kernel weights Kδ {ρ(η(z(θ)), η(y))} leads to the NP estimate of the posterior conditional density i ˜ Kb (θi − θ)Kδ {ρ(η(z(θi )), η(y))} i Kδ {ρ(η(z(θi )), η(y))} [Blum, JASA, 2010]
  67. 67. ABC-NP (density estimations) Other versions incorporating regression adjustments i ˜ Kb (θ∗ − θ)Kδ {ρ(η(z(θi )), η(y))} i i Kδ {ρ(η(z(θi )), η(y))} In all cases, error E[^ (θ|y)] − g (θ|y) = cb 2 + cδ2 + OP (b 2 + δ2 ) + OP (1/nδd ) g c var(^ (θ|y)) = g (1 + oP (1)) nbδd
  68. 68. ABC-NP (density estimations) Other versions incorporating regression adjustments i ˜ Kb (θ∗ − θ)Kδ {ρ(η(z(θi )), η(y))} i i Kδ {ρ(η(z(θi )), η(y))} In all cases, error E[^ (θ|y)] − g (θ|y) = cb 2 + cδ2 + OP (b 2 + δ2 ) + OP (1/nδd ) g c var(^ (θ|y)) = g (1 + oP (1)) nbδd [Blum, JASA, 2010]
  69. 69. ABC-NP (density estimations) Other versions incorporating regression adjustments i ˜ Kb (θ∗ − θ)Kδ {ρ(η(z(θi )), η(y))} i i Kδ {ρ(η(z(θi )), η(y))} In all cases, error E[^ (θ|y)] − g (θ|y) = cb 2 + cδ2 + OP (b 2 + δ2 ) + OP (1/nδd ) g c var(^ (θ|y)) = g (1 + oP (1)) nbδd [standard NP calculations]
  70. 70. ABC-NCH Incorporating non-linearities and heterocedasticities: σ(η(y)) ^ θ∗ = m(η(y)) + [θ − m(η(z))] ^ ^ σ(η(z)) ^ where m(η) estimated by non-linear regression (e.g., neural network) ^ σ(η) estimated by non-linear regression on residuals ^ log{θi − m(ηi )}2 = log σ2 (ηi ) + ξi ^ [Blum & Fran¸ois, 2009] c
  71. 71. ABC-NCH Incorporating non-linearities and heterocedasticities: σ(η(y)) ^ θ∗ = m(η(y)) + [θ − m(η(z))] ^ ^ σ(η(z)) ^ where m(η) estimated by non-linear regression (e.g., neural network) ^ σ(η) estimated by non-linear regression on residuals ^ log{θi − m(ηi )}2 = log σ2 (ηi ) + ξi ^ [Blum & Fran¸ois, 2009] c
  72. 72. ABC as knn Practice of ABC: determine tolerance as a quantile on observed distances, say 10% or 1% quantile, = N = qα (d1 , . . . , dN ) Interpretation of ε as non- parametric bandwidth only approximation of the actual practice [Blum & Fran¸ois, 2010] c ABC is a k-nearest neighbour (knn) method with kN = N N [Loftsgaarden & Quesenberry, 1965] [Biau et al., 2012, arxiv:1207.6461]
  73. 73. ABC as knn Practice of ABC: determine tolerance as a quantile on observed distances, say 10% or 1% quantile, = N = qα (d1 , . . . , dN ) Interpretation of ε as non- parametric bandwidth only approximation of the actual practice [Blum & Fran¸ois, 2010] c ABC is a k-nearest neighbour (knn) method with kN = N N [Loftsgaarden & Quesenberry, 1965] [Biau et al., 2012, arxiv:1207.6461]
  74. 74. ABC as knn Practice of ABC: determine tolerance as a quantile on observed distances, say 10% or 1% quantile, = N = qα (d1 , . . . , dN ) Interpretation of ε as non- parametric bandwidth only approximation of the actual practice [Blum & Fran¸ois, 2010] c ABC is a k-nearest neighbour (knn) method with kN = N N [Loftsgaarden & Quesenberry, 1965] [Biau et al., 2012, arxiv:1207.6461]
  75. 75. ABC consistency Provided kN / log log N −→ ∞ and kN /N −→ 0 as N → ∞, for almost all s0 (with respect to the distribution of S), with probability 1, kN 1 ϕ(θj ) −→ E[ϕ(θj )|S = s0 ] kN j=1 [Devroye, 1982] Biau et al. (2012) also recall pointwise and integrated mean square error consistency results on the corresponding kernel estimate of the conditional posterior distribution, under constraints p kN → ∞, kN /N → 0, hN → 0 and hN kN → ∞,
  76. 76. ABC consistency Provided kN / log log N −→ ∞ and kN /N −→ 0 as N → ∞, for almost all s0 (with respect to the distribution of S), with probability 1, kN 1 ϕ(θj ) −→ E[ϕ(θj )|S = s0 ] kN j=1 [Devroye, 1982] Biau et al. (2012) also recall pointwise and integrated mean square error consistency results on the corresponding kernel estimate of the conditional posterior distribution, under constraints p kN → ∞, kN /N → 0, hN → 0 and hN kN → ∞,
  77. 77. Rates of convergence Further assumptions (on target and kernel) allow for precise (integrated mean square) convergence rates (as a power of the sample size N), derived from classical k-nearest neighbour regression, like 4 when m = 1, 2, 3, kN ≈ N (p+4)/(p+8) and rate N − p+8 4 when m = 4, kN ≈ N (p+4)/(p+8) and rate N − p+8 log N 4 when m > 4, kN ≈ N (p+4)/(m+p+4) and rate N − m+p+4 [Biau et al., 2012, arxiv:1207.6461] Only applies to sufficient summary statistics
  78. 78. Rates of convergence Further assumptions (on target and kernel) allow for precise (integrated mean square) convergence rates (as a power of the sample size N), derived from classical k-nearest neighbour regression, like 4 when m = 1, 2, 3, kN ≈ N (p+4)/(p+8) and rate N − p+8 4 when m = 4, kN ≈ N (p+4)/(p+8) and rate N − p+8 log N 4 when m > 4, kN ≈ N (p+4)/(m+p+4) and rate N − m+p+4 [Biau et al., 2012, arxiv:1207.6461] Only applies to sufficient summary statistics
  79. 79. ABC inference machine Introduction ABC ABC as an inference machine Error inc. Exact BC and approximate targets summary statistic ABCel
  80. 80. How much Bayesian aBc is..? maybe a convergent method of inference (meaningful? sufficient? foreign?) approximation error unknown (w/o simulation) pragmatic Bayes (there is no other solution!) many calibration issues (tolerance, distance, statistics)
  81. 81. How much Bayesian aBc is..? maybe a convergent method of inference (meaningful? sufficient? foreign?) approximation error unknown (w/o simulation) pragmatic Bayes (there is no other solution!) many calibration issues (tolerance, distance, statistics) ...should Bayesians care?!
  82. 82. How much Bayesian aBc is..? maybe a convergent method of inference (meaningful? sufficient? foreign?) approximation error unknown (w/o simulation) pragmatic Bayes (there is no other solution!) many calibration issues (tolerance, distance, statistics) yes they should!!!
  83. 83. How much Bayesian aBc is..? maybe a convergent method of inference (meaningful? sufficient? foreign?) approximation error unknown (w/o simulation) pragmatic Bayes (there is no other solution!) many calibration issues (tolerance, distance, statistics) to ABCel
  84. 84. ABCµ Idea Infer about the error as well as about the parameter: Use of a joint density f (θ, |y) ∝ ξ( |y, θ) × πθ (θ) × π ( ) where y is the data, and ξ( |y, θ) is the prior predictive density of ρ(η(z), η(y)) given θ and y when z ∼ f (z|θ) Warning! Replacement of ξ( |y, θ) with a non-parametric kernel approximation. [Ratmann, Andrieu, Wiuf and Richardson, 2009, PNAS]
  85. 85. ABCµ Idea Infer about the error as well as about the parameter: Use of a joint density f (θ, |y) ∝ ξ( |y, θ) × πθ (θ) × π ( ) where y is the data, and ξ( |y, θ) is the prior predictive density of ρ(η(z), η(y)) given θ and y when z ∼ f (z|θ) Warning! Replacement of ξ( |y, θ) with a non-parametric kernel approximation. [Ratmann, Andrieu, Wiuf and Richardson, 2009, PNAS]
  86. 86. ABCµ Idea Infer about the error as well as about the parameter: Use of a joint density f (θ, |y) ∝ ξ( |y, θ) × πθ (θ) × π ( ) where y is the data, and ξ( |y, θ) is the prior predictive density of ρ(η(z), η(y)) given θ and y when z ∼ f (z|θ) Warning! Replacement of ξ( |y, θ) with a non-parametric kernel approximation. [Ratmann, Andrieu, Wiuf and Richardson, 2009, PNAS]
  87. 87. ABCµ details Multidimensional distances ρk (k = 1, . . . , K ) and errors k = ρk (ηk (z), ηk (y)), with 1 k ∼ ξk ( |y, θ) ≈ ξk ( |y, θ) = ^ K [{ k −ρk (ηk (zb ), ηk (y))}/hk ] Bhk b then used in replacing ξ( |y, θ) with mink ξk ( |y, θ) ^ ABCµ involves acceptance probability π(θ , ) q(θ , θ)q( , ) mink ξk ( |y, θ ) ^ π(θ, ) q(θ, θ )q( , ) mink ξk ( |y, θ) ^
  88. 88. ABCµ details Multidimensional distances ρk (k = 1, . . . , K ) and errors k = ρk (ηk (z), ηk (y)), with 1 k ∼ ξk ( |y, θ) ≈ ξk ( |y, θ) = ^ K [{ k −ρk (ηk (zb ), ηk (y))}/hk ] Bhk b then used in replacing ξ( |y, θ) with mink ξk ( |y, θ) ^ ABCµ involves acceptance probability π(θ , ) q(θ , θ)q( , ) mink ξk ( |y, θ ) ^ π(θ, ) q(θ, θ )q( , ) mink ξk ( |y, θ) ^
  89. 89. ABCµ multiple errors [ c Ratmann et al., PNAS, 2009]
  90. 90. ABCµ for model choice [ c Ratmann et al., PNAS, 2009]
  91. 91. Wilkinson’s exact BC (not exactly!) ABC approximation error (i.e. non-zero tolerance) replaced with exact simulation from a controlled approximation to the target, convolution of true posterior with kernel function π(θ)f (z|θ)K (y − z) π (θ, z|y) = , π(θ)f (z|θ)K (y − z)dzdθ with K kernel parameterised by bandwidth . [Wilkinson, 2008] Theorem The ABC algorithm based on the assumption of a randomised observation y = y + ξ, ξ ∼ K , and an acceptance probability of ˜ K (y − z)/M gives draws from the posterior distribution π(θ|y).
  92. 92. Wilkinson’s exact BC (not exactly!) ABC approximation error (i.e. non-zero tolerance) replaced with exact simulation from a controlled approximation to the target, convolution of true posterior with kernel function π(θ)f (z|θ)K (y − z) π (θ, z|y) = , π(θ)f (z|θ)K (y − z)dzdθ with K kernel parameterised by bandwidth . [Wilkinson, 2008] Theorem The ABC algorithm based on the assumption of a randomised observation y = y + ξ, ξ ∼ K , and an acceptance probability of ˜ K (y − z)/M gives draws from the posterior distribution π(θ|y).
  93. 93. How exact a BC? “Using to represent measurement error is straightforward, whereas using to model the model discrepancy is harder to conceptualize and not as commonly used” [Richard Wilkinson, 2008]
  94. 94. How exact a BC? Pros Pseudo-data from true model and observed data from noisy model Interesting perspective in that outcome is completely controlled Link with ABCµ and assuming y is observed with a measurement error with density K Relates to the theory of model approximation [Kennedy & O’Hagan, 2001] Cons Requires K to be bounded by M True approximation error never assessed Requires a modification of the standard ABC algorithm
  95. 95. ABC for HMMs Specific case of a hidden Markov model Xt+1 ∼ Qθ (Xt , ·) Yt+1 ∼ gθ (·|xt ) where only y0 is observed. 1:n [Dean, Singh, Jasra, & Peters, 2011] Use of specific constraints, adapted to the Markov structure: y1 ∈ B(y1 , ) × · · · × yn ∈ B(yn , ) 0 0
  96. 96. ABC for HMMs Specific case of a hidden Markov model Xt+1 ∼ Qθ (Xt , ·) Yt+1 ∼ gθ (·|xt ) where only y0 is observed. 1:n [Dean, Singh, Jasra, & Peters, 2011] Use of specific constraints, adapted to the Markov structure: y1 ∈ B(y1 , ) × · · · × yn ∈ B(yn , ) 0 0
  97. 97. ABC-MLE for HMMs ABC-MLE defined by θn = arg max Pθ Y1 ∈ B(y1 , ), . . . , Yn ∈ B(yn , ) ^ 0 0 θ Exact MLE for the likelihood same basis as Wilkinson! 0 pθ (y1 , . . . , yn ) corresponding to the perturbed process (xt , yt + zt )1 t n zt ∼ U(B(0, 1) [Dean, Singh, Jasra, & Peters, 2011]
  98. 98. ABC-MLE for HMMs ABC-MLE defined by θn = arg max Pθ Y1 ∈ B(y1 , ), . . . , Yn ∈ B(yn , ) ^ 0 0 θ Exact MLE for the likelihood same basis as Wilkinson! 0 pθ (y1 , . . . , yn ) corresponding to the perturbed process (xt , yt + zt )1 t n zt ∼ U(B(0, 1) [Dean, Singh, Jasra, & Peters, 2011]
  99. 99. ABC-MLE is biased ABC-MLE is asymptotically (in n) biased with target l (θ) = Eθ∗ [log pθ (Y1 |Y−∞:0 )] but ABC-MLE converges to true value in the sense l n (θn ) → l (θ) for all sequences (θn ) converging to θ and n
  100. 100. ABC-MLE is biased ABC-MLE is asymptotically (in n) biased with target l (θ) = Eθ∗ [log pθ (Y1 |Y−∞:0 )] but ABC-MLE converges to true value in the sense l n (θn ) → l (θ) for all sequences (θn ) converging to θ and n
  101. 101. Noisy ABC-MLE Idea: Modify instead the data from the start 0 (y1 + ζ1 , . . . , yn + ζn ) [ see Fearnhead-Prangle ] noisy ABC-MLE estimate arg max Pθ Y1 ∈ B(y1 + ζ1 , ), . . . , Yn ∈ B(yn + ζn , ) 0 0 θ [Dean, Singh, Jasra, & Peters, 2011]
  102. 102. Consistent noisy ABC-MLE Degrading the data improves the estimation performances: Noisy ABC-MLE is asymptotically (in n) consistent under further assumptions, the noisy ABC-MLE is asymptotically normal increase in variance of order −2 likely degradation in precision or computing time due to the lack of summary statistic [curse of dimensionality]
  103. 103. SMC for ABC likelihood Algorithm 2 SMC ABC for HMMs Given θ for k = 1, . . . , n do 1 1 N N generate proposals (xk , yk ), . . . , (xk , yk ) from the model weigh each proposal with ωk l =I l B(yk + ζk , ) (yk ) 0 l renormalise the weights and sample the xk ’s accordingly end for approximate the likelihood by n N ωlk N k=1 l=1 [Jasra, Singh, Martin, & McCoy, 2010]
  104. 104. Which summary? Fundamental difficulty of the choice of the summary statistic when there is no non-trivial sufficient statistics Starting from a large collection of summary statistics is available, Joyce and Marjoram (2008) consider the sequential inclusion into the ABC target, with a stopping rule based on a likelihood ratio test Not taking into account the sequential nature of the tests Depends on parameterisation Order of inclusion matters likelihood ratio test?!
  105. 105. Which summary? Fundamental difficulty of the choice of the summary statistic when there is no non-trivial sufficient statistics Starting from a large collection of summary statistics is available, Joyce and Marjoram (2008) consider the sequential inclusion into the ABC target, with a stopping rule based on a likelihood ratio test Not taking into account the sequential nature of the tests Depends on parameterisation Order of inclusion matters likelihood ratio test?!
  106. 106. Which summary? Fundamental difficulty of the choice of the summary statistic when there is no non-trivial sufficient statistics Starting from a large collection of summary statistics is available, Joyce and Marjoram (2008) consider the sequential inclusion into the ABC target, with a stopping rule based on a likelihood ratio test Not taking into account the sequential nature of the tests Depends on parameterisation Order of inclusion matters likelihood ratio test?!
  107. 107. Which summary for model choice? Depending on the choice of η(·), the Bayes factor based on this insufficient statistic, η π1 (θ1 )f1η (η(y)|θ1 ) dθ1 B12 (y) = , π2 (θ2 )f2η (η(y)|θ2 ) dθ2 is consistent or not. [X, Cornuet, Marin, & Pillai, 2012] Consistency only depends on the range of Ei [η(y)] under both models. [Marin, Pillai, X, & Rousseau, 2012]
  108. 108. Which summary for model choice? Depending on the choice of η(·), the Bayes factor based on this insufficient statistic, η π1 (θ1 )f1η (η(y)|θ1 ) dθ1 B12 (y) = , π2 (θ2 )f2η (η(y)|θ2 ) dθ2 is consistent or not. [X, Cornuet, Marin, & Pillai, 2012] Consistency only depends on the range of Ei [η(y)] under both models. [Marin, Pillai, X, & Rousseau, 2012]
  109. 109. Semi-automatic ABC Fearnhead and Prangle (2010) study ABC and the selection of the summary statistic in close proximity to Wilkinson’s proposal ABC considered as inferential method and calibrated as such randomised (or ‘noisy’) version of the summary statistics ˜ η(y) = η(y) + τ derivation of a well-calibrated version of ABC, i.e. an algorithm that gives proper predictions for the distribution associated with this randomised summary statistic
  110. 110. Summary [of F&P/statistics) optimality of the posterior expectation E[θ|y] of the parameter of interest as summary statistics η(y)! use of the standard quadratic loss function (θ − θ0 )T A(θ − θ0 ) . recent extension to model choice, optimality of Bayes factor B12 (y) [F&P, ISBA 2012 talk]
  111. 111. Summary [of F&P/statistics) optimality of the posterior expectation E[θ|y] of the parameter of interest as summary statistics η(y)! use of the standard quadratic loss function (θ − θ0 )T A(θ − θ0 ) . recent extension to model choice, optimality of Bayes factor B12 (y) [F&P, ISBA 2012 talk]
  112. 112. Conclusion Choice of summary statistics is paramount for ABC validation/performance At best, ABC approximates π(. | η(y)) Model selection feasible with ABC [with caution!] For estimation, consistency if {θ; µ(θ) = µ0 } = θ0 For testing consistency if {µ1 (θ1 ), θ1 ∈ Θ1 } ∩ {µ2 (θ2 ), θ2 ∈ Θ2 } = ∅ [Marin et al., 2011]
  113. 113. Conclusion Choice of summary statistics is paramount for ABC validation/performance At best, ABC approximates π(. | η(y)) Model selection feasible with ABC [with caution!] For estimation, consistency if {θ; µ(θ) = µ0 } = θ0 For testing consistency if {µ1 (θ1 ), θ1 ∈ Θ1 } ∩ {µ2 (θ2 ), θ2 ∈ Θ2 } = ∅ [Marin et al., 2011]
  114. 114. Conclusion Choice of summary statistics is paramount for ABC validation/performance At best, ABC approximates π(. | η(y)) Model selection feasible with ABC [with caution!] For estimation, consistency if {θ; µ(θ) = µ0 } = θ0 For testing consistency if {µ1 (θ1 ), θ1 ∈ Θ1 } ∩ {µ2 (θ2 ), θ2 ∈ Θ2 } = ∅ [Marin et al., 2011]
  115. 115. Conclusion Choice of summary statistics is paramount for ABC validation/performance At best, ABC approximates π(. | η(y)) Model selection feasible with ABC [with caution!] For estimation, consistency if {θ; µ(θ) = µ0 } = θ0 For testing consistency if {µ1 (θ1 ), θ1 ∈ Θ1 } ∩ {µ2 (θ2 ), θ2 ∈ Θ2 } = ∅ [Marin et al., 2011]
  116. 116. Conclusion Choice of summary statistics is paramount for ABC validation/performance At best, ABC approximates π(. | η(y)) Model selection feasible with ABC [with caution!] For estimation, consistency if {θ; µ(θ) = µ0 } = θ0 For testing consistency if {µ1 (θ1 ), θ1 ∈ Θ1 } ∩ {µ2 (θ2 ), θ2 ∈ Θ2 } = ∅ [Marin et al., 2011]
  117. 117. Empirical likelihood (EL) Introduction ABC ABC as an inference machine ABCel ABC and EL Composite likelihood Illustrations
  118. 118. Empirical likelihood (EL) Dataset x made of n independent replicates x = (x1 , . . . , xn ) of some X ∼ F Generalized moment condition model EF h(X , φ) = 0, where h is a known function, and φ an unknown parameter Corresponding empirical likelihood n Lel (φ|x) = max pi p i=1 for all p such that 0 pi 1, pi = 1, i pi h(xi , φ) = 0. [Owen, 1988, Bio’ka; Owen, 2001]
  119. 119. Empirical likelihood (EL) Dataset x made of n independent replicates x = (x1 , . . . , xn ) of some X ∼ F Generalized moment condition model EF h(X , φ) = 0, where h is a known function, and φ an unknown parameter Corresponding empirical likelihood n Lel (φ|x) = max pi p i=1 for all p such that 0 pi 1, pi = 1, i pi h(xi , φ) = 0. [Owen, 1988, Bio’ka; Owen, 2001]
  120. 120. Convergence of EL [3.4] Theorem 3.4 Let X , Y1 , . . . , Yn be independent rv’s with common distribution F0 . For θ ∈ Θ, and the function h(X , θ) ∈ Rs , let θ0 ∈ Θ be such that Var(h(Yi , θ0 )) is finite and has rank q > 0. If θ0 satisfies E(h(X , θ0 )) = 0, then Lel (θ0 |Y1 , . . . , Yn ) −2 log → χ2 (q) n−n in distribution when n → ∞. [Owen, 2001]
  121. 121. Convergence of EL [3.4] “...The interesting thing about Theorem 3.4 is what is not there. It ^ includes no conditions to make θ a good estimate of θ0 , nor even conditions to ensure a unique value for θ0 , nor even that any solution θ0 exists. Theorem 3.4 applies in the just determined, over-determined, and under-determined cases. When we can prove that our estimating ^ equations uniquely define θ0 , and provide a consistent estimator θ of it, then confidence regions and tests follow almost automatically through Theorem 3.4.”. [Owen, 2001]
  122. 122. Raw ABCel sampler We act as if EL was an exact likelihood for i = 1 → N do generate φi from the prior distribution π(·) set the weight ωi = Lel (φi |xobs ) end for return (φi , ωi ), i = 1, . . . , N The output is sample of parameters of size N with associated weights [Cornuet et al., 2012]
  123. 123. Raw ABCel sampler We act as if EL was an exact likelihood for i = 1 → N do generate φi from the prior distribution π(·) set the weight ωi = Lel (φi |xobs ) end for return (φi , ωi ), i = 1, . . . , N Performance of the output evaluated through effective sample size  2 N  N  ESS = 1 ωi ωj   i=1 j=1 [Cornuet et al., 2012]
  124. 124. Raw ABCel sampler We act as if EL was an exact likelihood for i = 1 → N do generate φi from the prior distribution π(·) set the weight ωi = Lel (φi |xobs ) end for return (φi , ωi ), i = 1, . . . , N Other classical sampling algorithms might be adapted to use EL. We resorted to the adaptive multiple importance sampling (AMIS) of Cornuet to speed up computations [Cornuet et al., 2012]
  125. 125. Moment condition in population genetics? EL does not require a fully defined and often complex (hence debatable) parametric model Main difficulty Derive a constraint EF h(X , φ) = 0, on the parameters of interest φ when X is made of the genotypes of the sample of individuals at a given locus E.g., in phylogeography, φ is composed of dates of divergence between populations, ratio of population sizes, mutation rates, etc. None of them are moments of the distribution of the allelic states of the sample
  126. 126. Moment condition in population genetics? EL does not require a fully defined and often complex (hence debatable) parametric model Main difficulty Derive a constraint EF h(X , φ) = 0, on the parameters of interest φ when X is made of the genotypes of the sample of individuals at a given locus c h pairwise composite scores whose zero is the pairwise maximum likelihood estimator
  127. 127. Pairwise composite likelihood The intra-locus pairwise likelihood j 2 (xk |φ) 2 (xk , xk |φ) i = i<j 1 n with xk , . . . , xk : allelic states of the gene sample at the k-th locus The pairwise score function j φ log 2 (xk |φ) φ log 2 (xk , xk |φ) i = i<j Composite likelihoods are often much narrower than the original likelihood of the model Safe with EL because we only use position of its mode
  128. 128. Pairwise likelihood: a simple case 1 Assumptions 2 (δ|θ) =√ ρ (θ)|δ| 1 + 2θ sample ⊂ closed, panmictic with θ population at equilibrium ρ(θ) = √ 1 + θ + 1 + 2θ marker: microsatellite mutation rate: θ/2 Pairwise score function ∂θ log 2 (δ|θ) = i j 1 |δ| if xk et xk are two genes of the − + √ sample, 1 + 2θ θ 1 + 2θ j 2 (xk , xk |θ) i depends only on i j δ = xk − xk
  129. 129. Pairwise likelihood: a simple case 1 Assumptions 2 (δ|θ) =√ ρ (θ)|δ| 1 + 2θ sample ⊂ closed, panmictic with θ population at equilibrium ρ(θ) = √ 1 + θ + 1 + 2θ marker: microsatellite mutation rate: θ/2 Pairwise score function ∂θ log 2 (δ|θ) = i j 1 |δ| if xk et xk are two genes of the − + √ sample, 1 + 2θ θ 1 + 2θ j 2 (xk , xk |θ) i depends only on i j δ = xk − xk
  130. 130. Pairwise likelihood: a simple case 1 Assumptions 2 (δ|θ) =√ ρ (θ)|δ| 1 + 2θ sample ⊂ closed, panmictic with θ population at equilibrium ρ(θ) = √ 1 + θ + 1 + 2θ marker: microsatellite mutation rate: θ/2 Pairwise score function ∂θ log 2 (δ|θ) = i j 1 |δ| if xk et xk are two genes of the − + √ sample, 1 + 2θ θ 1 + 2θ j 2 (xk , xk |θ) i depends only on i j δ = xk − xk
  131. 131. Pairwise likelihood: 2 diverging populations MRCA Then 2 (δ|θ, τ) = +∞ e−τθ √ ρ(θ)|k| Iδ−k (τθ). τ 1 + 2θ k=−∞ where In (z) nth-order modified POP a POP b Bessel function of the first Assumptions kind τ: divergence date of pop. a and b θ/2: mutation rate i j Let xk and xk be two genes coming resp. from pop. a and b i j Set δ = xk − xk .
  132. 132. Pairwise likelihood: 2 diverging populations MRCA A 2-dim score function ∂τ log 2 (δ|θ, τ) = τ −θ + θ 2 (δ − 1|θ, τ) + 2 (δ + 1|θ, τ) 2 2 (δ|θ, τ) POP a POP b ∂θ log 2 (δ|θ, τ) = Assumptions 1 −τ − + 1 + 2θ τ: divergence date of τ 2 (δ − 1|θ, τ) + 2 (δ + 1|θ, τ) + pop. a and b 2 2 (δ|θ, τ) q(δ|θ, τ) θ/2: mutation rate 2 (δ|θ, τ) i j Let xk and xk be two genes coming resp. from pop. a and where b q(δ|θ, τ) := ∞ i j Set δ = xk − xk . e−τθ ρ (θ) √ |k|ρ(θ)|k| Iδ−k (τθ) 1 + 2θ ρ(θ) k=−∞
  133. 133. Example: normal posterior ABCel with two constraints ESS=108.9 ESS=81.48 ESS=105.2 3.0 2.0 2.0 Density Density Density 1.5 1.0 1.0 0.0 0.0 0.0 −0.4 0.0 0.2 0.4 0.6 −0.8 −0.6 −0.4 −0.2 0.0 −0.2 0.0 0.2 0.4 0.6 0.8 ESS=133.3 θ ESS=87.75 θ ESS=72.89 θ 3.0 2.0 4 2.0 Density Density Density 1.0 1.0 2 0.0 0.0 0 −0.8 −0.4 0.0 0.2 0.4 −0.2 0.0 0.2 0.4 0.6 0.8 −0.2 0.0 0.2 0.4 ESS=116.5 θ ESS=103.9 θ ESS=126.9 θ 3.0 2.0 3.0 2.0 Density Density Density 1.0 1.5 1.0 0.0 0.0 −0.4 0.0 0.2 0.4 0.6 −0.4 −0.2 0.0 0.2 0.4 0.0 −0.8 −0.4 0.0 0.4 ESS=113.3 θ ESS=92.99 θ ESS=121.4 θ 3.0 2.0 2.0 Density Density Density 1.5 1.0 1.0 0.0 0.0 0.0 −0.6 −0.2 0.2 0.6 −0.2 0.0 0.2 0.4 0.6 −0.5 0.0 0.5 ESS=133.6 θ ESS=116.4 θ ESS=131.6 θ 0.0 1.0 2.0 3.0 2.0 Density Density Density 1.0 1.0 0.0 0.0 −0.6 −0.2 0.2 0.4 0.6 −0.4 −0.2 0.0 0.2 0.4 −0.5 0.0 0.5 Sample sizes are of 21 (column 3), 41 (column 1) and 61 (column 2) observations

×