Rename PriorAbilityEstimator => PosteriorAbilityEstimator

Author: Frankie Robertson

benchmark/benchmarks.jl

@@ -30,7 +30,7 @@ function prepare_4pls(group)
     integrator = even_grid(-6.0, 6.0, mirtcat_quadpts(1))
     optimizer = AbilityOptimizer(OneDimOptimOptimizer(-6.0, 6.0, NelderMead()))
 
-    dist_ability_estimator = PriorAbilityEstimator()
+    dist_ability_estimator = PosteriorAbilityEstimator()
     ability_estimators = [
         ("mean", MeanAbilityEstimator(dist_ability_estimator, integrator)),
         ("mode", ModeAbilityEstimator(dist_ability_estimator, optimizer))

docs/examples/examples/ability_convergence_3pl.jl

@@ -22,7 +22,7 @@ using ComputerAdaptiveTesting
 using ComputerAdaptiveTesting.Sim: auto_responder
 using ComputerAdaptiveTesting.NextItemRules: AbilityVarianceStateCriterion
 using ComputerAdaptiveTesting.TerminationConditions: FixedItemsTerminationCondition
-using ComputerAdaptiveTesting.Aggregators: PriorAbilityEstimator,
+using ComputerAdaptiveTesting.Aggregators: PosteriorAbilityEstimator,
     MeanAbilityEstimator, LikelihoodAbilityEstimator
 using FittedItemBanks
 using ComputerAdaptiveTesting.Responses: BooleanResponse
@@ -46,7 +46,7 @@ using FittedItemBanks.DummyData: dummy_full, std_normal, SimpleItemBankSpec, Std
 # CatRecorder collects information which can be used to draw different types of plots.
 max_questions = 99
 integrator = FixedGKIntegrator(-6, 6, 80)
-dist_ability_est = PriorAbilityEstimator(std_normal)
+dist_ability_est = PosteriorAbilityEstimator(std_normal)
 ability_estimator = MeanAbilityEstimator(dist_ability_est, integrator)
 rules = CatRules(ability_estimator,
     AbilityVarianceStateCriterion(dist_ability_est, integrator),

docs/examples/examples/ability_convergence_mirt.jl

@@ -22,7 +22,7 @@ using ComputerAdaptiveTesting
 using ComputerAdaptiveTesting.Sim: auto_responder
 using ComputerAdaptiveTesting.NextItemRules: DRuleItemCriterion
 using ComputerAdaptiveTesting.TerminationConditions: FixedItemsTerminationCondition
-using ComputerAdaptiveTesting.Aggregators: PriorAbilityEstimator,
+using ComputerAdaptiveTesting.Aggregators: PosteriorAbilityEstimator,
     MeanAbilityEstimator, LikelihoodAbilityEstimator
 using FittedItemBanks
 import PsychometricsBazaarBase.IntegralCoeffs
@@ -49,7 +49,7 @@ using ComputerAdaptiveTesting.Responses: BooleanResponse
 # CatRecorder collects information which can be used to draw different types of plots.
 max_questions = 9
 integrator = CubaIntegrator([-6.0, -6.0], [6.0, 6.0], CubaVegas(); rtol = 1e-2)
-ability_estimator = MeanAbilityEstimator(PriorAbilityEstimator(std_mv_normal(dims)),
+ability_estimator = MeanAbilityEstimator(PosteriorAbilityEstimator(std_mv_normal(dims)),
     integrator)
 rules = CatRules(ability_estimator,
     DRuleItemCriterion(ability_estimator),

docs/examples/examples/vocab_iq.jl

@@ -5,9 +5,9 @@
 # ---
 
 #md # Running a CAT based based on real response data
-# 
+#
 # This example shows how to run a CAT end-to-end on real data.
-# 
+#
 # First a 1-dimensional IRT model is fitted based on open response data to the
 # vocabulary IQ test using the IRTSupport package which internally, this uses
 # the `mirt` R package. Next, the model is used to administer the test
@@ -37,7 +37,7 @@ function run_vocab_iq_cat()
     item_bank, labels = get_item_bank()
     integrator = FixedGKIntegrator(-6, 6, 61)
     ability_integrator = AbilityIntegrator(integrator)
-    dist_ability_est = PriorAbilityEstimator(std_normal)
+    dist_ability_est = PosteriorAbilityEstimator(std_normal)
     optimizer = AbilityOptimizer(OneDimOptimOptimizer(-6.0, 6.0, NelderMead()))
     ability_estimator = ModeAbilityEstimator(dist_ability_est, optimizer)
     @info "run_cat" ability_estimator

profile/next_items.jl

@@ -20,7 +20,7 @@ function get_ability_estimator(multidim)
         integrator = FixedGKIntegrator(-6.0, 6.0)
         dist = Normal()
     end
-    return PriorAbilityEstimator(dist, integrator)
+    return PosteriorAbilityEstimator(dist, integrator)
 end
 
 function prepare_empty(item_bank, actual_responses, ability_tracker)

src/Compat/CatR.jl

@@ -5,7 +5,7 @@ using ComputerAdaptiveTesting.Aggregators: AbilityIntegrator,
                                            DistributionAbilityEstimator,
                                            ModeAbilityEstimator,
                                            MeanAbilityEstimator,
-                                           PriorAbilityEstimator
+                                           PosteriorAbilityEstimator
 using ComputerAdaptiveTesting.TerminationConditions: RunForeverTerminationCondition
 using ComputerAdaptiveTesting.Rules: CatRules
 using ComputerAdaptiveTesting.NextItemRules
@@ -51,9 +51,9 @@ const next_item_aliases = _next_item_aliases()
 
 function _ability_estimator_aliases()
     res = Dict{String, Any}()
-    res["BM"] = (; optimizer, kwargs...) -> ModeAbilityEstimator(PriorAbilityEstimator(), optimizer)
+    res["BM"] = (; optimizer, kwargs...) -> ModeAbilityEstimator(PosteriorAbilityEstimator(), optimizer)
     res["ML"] = (; optimizer, kwargs...) -> ModeAbilityEstimator(LikelihoodAbilityEstimator(), optimizer)
-    res["EAP"] = (; integrator, kwargs...) -> MeanAbilityEstimator(PriorAbilityEstimator(), integrator)
+    res["EAP"] = (; integrator, kwargs...) -> MeanAbilityEstimator(PosteriorAbilityEstimator(), integrator)
     #res["WL"]
     #res["ROB"]
     return res
@@ -97,7 +97,7 @@ function assemble_rules(;
     integrator = setup_integrator()
     optimizer = setup_optimizer()
     ability_estimator = ability_estimator_aliases[method](; integrator, optimizer)
-    posterior_ability_estimator = PriorAbilityEstimator()
+    posterior_ability_estimator = PosteriorAbilityEstimator()
     raw_next_item = next_item_aliases[criterion](ability_estimator, integrator, optimizer; posterior_ability_estimator=posterior_ability_estimator)
     next_item = FixedFirstItemNextItemRule(start_item, raw_next_item)
     CatRules(;
@@ -108,4 +108,4 @@ function assemble_rules(;
     )
 end
 
-end
+end

src/Compat/MirtCAT.jl

@@ -5,7 +5,7 @@ using ComputerAdaptiveTesting.Aggregators: SafeLikelihoodAbilityEstimator,
                                            DistributionAbilityEstimator,
                                            ModeAbilityEstimator,
                                            MeanAbilityEstimator,
-                                           PriorAbilityEstimator,
+                                           PosteriorAbilityEstimator,
                                            AbilityEstimator,
                                            distribution_estimator
 using ComputerAdaptiveTesting.TerminationConditions: RunForeverTerminationCondition
@@ -71,9 +71,9 @@ to randomly select items, and 'seq' for selecting items sequentially
 =#
 
 const ability_estimator_aliases = Dict(
-    "MAP" => (; optimizer, ncomp, kwargs...) -> ModeAbilityEstimator(PriorAbilityEstimator(; ncomp=ncomp), optimizer),
+    "MAP" => (; optimizer, ncomp, kwargs...) -> ModeAbilityEstimator(PosteriorAbilityEstimator(; ncomp=ncomp), optimizer),
     "ML" => (; optimizer, ncomp, kwargs...) -> ModeAbilityEstimator(SafeLikelihoodAbilityEstimator(; ncomp=ncomp), optimizer),
-    "EAP" => (; integrator, ncomp, kwargs...) -> MeanAbilityEstimator(PriorAbilityEstimator(; ncomp=ncomp), integrator),
+    "EAP" => (; integrator, ncomp, kwargs...) -> MeanAbilityEstimator(PosteriorAbilityEstimator(; ncomp=ncomp), integrator),
 # "WLE" for weighted likelihood estimation
 # "EAPsum" for the expected a-posteriori for each sum score
 )
@@ -143,7 +143,7 @@ function assemble_rules(;
     integrator = setup_integrator(lo, hi, pts)
     optimizer = setup_optimizer(-theta_lim, theta_lim)
     ability_estimator = ability_estimator_aliases[method](; integrator, optimizer, ncomp)
-    posterior_ability_estimator = PriorAbilityEstimator(; ncomp)
+    posterior_ability_estimator = PosteriorAbilityEstimator(; ncomp)
     raw_next_item = next_item_aliases[criteria](ability_estimator, posterior_ability_estimator, integrator, optimizer)
     next_item = FixedFirstItemNextItemRule(start_item, raw_next_item)
     CatRules(;
@@ -153,4 +153,4 @@ function assemble_rules(;
     )
 end
 
-end
+end

src/aggregators/Aggregators.jl

@@ -42,7 +42,7 @@ export AbilityEstimator, TrackedResponses
 export AbilityTracker, NullAbilityTracker, PointAbilityTracker, GriddedAbilityTracker
 export ClosedFormNormalAbilityTracker, track!
 export response_expectation, expectation, distribution_estimator
-export PointAbilityEstimator, PriorAbilityEstimator
+export PointAbilityEstimator, PosteriorAbilityEstimator
 export SafeLikelihoodAbilityEstimator, LikelihoodAbilityEstimator
 export ModeAbilityEstimator, MeanAbilityEstimator
 export Speculator, replace_speculation!, normdenom, maybe_tracked_ability_estimate

src/aggregators/ability_estimator.jl

@@ -26,19 +26,19 @@ function pdf(::LikelihoodAbilityEstimator,
     AbilityLikelihood(tracked_responses)
 end
 
-struct PriorAbilityEstimator{PriorT <: Distribution} <: DistributionAbilityEstimator
+struct PosteriorAbilityEstimator{PriorT <: Distribution} <: DistributionAbilityEstimator
     prior::PriorT
 end
 
-function PriorAbilityEstimator(; ncomp = 0)
+function PosteriorAbilityEstimator(; ncomp = 0)
     if ncomp == 0
-        return PriorAbilityEstimator(std_normal)
+        return PosteriorAbilityEstimator(std_normal)
     else
-        return PriorAbilityEstimator(std_mv_normal(ncomp))
+        return PosteriorAbilityEstimator(std_mv_normal(ncomp))
     end
 end
 
-function pdf(est::PriorAbilityEstimator,
+function pdf(est::PosteriorAbilityEstimator,
         tracked_responses::TrackedResponses)
     IntegralCoeffs.PriorApply(IntegralCoeffs.Prior(est.prior),
         AbilityLikelihood(tracked_responses))
@@ -75,7 +75,7 @@ end
 function SafeLikelihoodAbilityEstimator(args...; kwargs...)
     GuardedAbilityEstimator(
         LikelihoodAbilityEstimator(),
-        PriorAbilityEstimator(args...),
+        PosteriorAbilityEstimator(args...),
         multiple_response_types_guard
     )
 end
@@ -289,7 +289,7 @@ function (est::MeanAbilityEstimator{AbilityEstimatorT, RiemannEnumerationIntegra
         tracked_responses)
 end
 
-function maybe_apply_prior(f::F, est::PriorAbilityEstimator) where {F}
+function maybe_apply_prior(f::F, est::PosteriorAbilityEstimator) where {F}
     IntegralCoeffs.PriorApply(IntegralCoeffs.Prior(est.prior), f)
 end
 

src/aggregators/ability_trackers/closed_form_normal.jl

@@ -2,7 +2,7 @@ mutable struct ClosedFormNormalAbilityTracker <: AbilityTracker
     cur_ability::VarNormal
 end
 
-function ClosedFormNormalAbilityTracker(prior_ability_estimator::PriorAbilityEstimator)
+function ClosedFormNormalAbilityTracker(prior_ability_estimator::PosteriorAbilityEstimator)
     @warn "ClosedFormNormalAbilityTracker is based on equations from Liden 1998 / Owen 1975, but these appear to give poor results"
     prior = prior_ability_estimator.prior
     if !(prior isa Normal)