Package 'jheem2'

Description

Subset an ont

Usage

## S3 method for class 'ontology'
ont[i]

Arguments

Value

An ontology

Description

Modify an ontology

Usage

## S3 replacement method for class 'ontology'
ont[[i]] <- value

Arguments

Value

An ontology

Description

Modify an ontology

Usage

## S3 replacement method for class 'ontology'
ont[i] <- value

Arguments

Value

An ontology

Description

Modify an ontology

Usage

## S3 replacement method for class 'ontology'
ont$i <- value

Arguments

Value

An ontology

Description

Add Sub Version to Specification

Usage

add.sub.version(
  specification,
  sub.version,
  description,
  inherit.outcomes = T,
  can.seed.new.engine = T
)

Description

Anchor Intervention Effect's Location and Version

Usage

anchor.intervention.effect.location.and.version(
  intervention.effect,
  location,
  specification.metadata,
  error.prefix = ""
)

Arguments

Value

A new intervention effect that has location and version set

Description

Apply an Ontology Mapping to Data

Usage

apply.ontology.mapping(
  ontology.mapping,
  from.arr,
  to.dim.names = NULL,
  fun = "sum",
  na.rm = F,
  error.prefix = ""
)

Arguments

See Also

Other Functions to identify why ontologies fail to map: get.identity.ontology.mapping(), map.value.ontology()

Description

Summarize the progress of a calibration task

Usage

are.simulations.present(
  version,
  locations,
  calibration.code,
  interventions,
  n.sim,
  sub.version = NULL,
  root.dir = get.jheem.root.directory()
)

Arguments

Description

Flexibly Subset an Array

Usage

array.access(arr, ..., drop = F)

fast.array.access(arr, dimension.values, drop = F)

Arguments

Functions

• fast.array.access():

Examples

arr = array(1:24, dim=c(age=4, race=3, sex=2), 
                 dimnames=list(age=c('age1','age2','age3','age4'), 
                               race=c('black','hispanic','other'), 
                               sex=c('female','male')))
                               
array.access(arr, dimension.values=list(age=1:3, race='black'))
#is equivalent to
arr[1:3, 'black',]

Description

Convert an ontology back to a plain list

Usage

## S3 method for class 'ontology'
as.list(ont)

Arguments

Description

Convert a named list to an ontology

Usage

as.ontology(x, incomplete.dimensions = NULL)

Arguments

Description

Assemble an MCMC From a Calibration

Usage

assemble.mcmc.from.calibration(
  version,
  location,
  calibration.code,
  root.dir = get.jheem.root.directory("Cannot set up calibration: "),
  allow.incomplete = F,
  chains = NULL
)

Arguments

Description

Assemble Simulations From A Calibration

Usage

assemble.simulations.from.calibration(
  version,
  location,
  calibration.code,
  root.dir = get.jheem.root.directory("Cannot set up calibration: "),
  allow.incomplete = F,
  chains = NULL,
  verbose = T
)

Arguments

Description

Assemble Transmuted Simulations

Usage

assemble.transmuted.simulations(
  transmute.code,
  location,
  n.sim,
  to.version = NULL,
  to.sub.version = NULL,
  chunks = NULL,
  allow.incomplete = F,
  root.dir = get.jheem.root.directory()
)

Description

Subset a simulation set by trimming off a portion of earlier simulations.

Usage

burn.simset(simset, keep = NULL)

Arguments

Details

Decimal values above 1 are rounded down to the nearest whole number.

Description

Cache A Summary of an MCMC

Usage

cache.mcmc.summary(
  version,
  location,
  calibration.code,
  root.dir = get.jheem.root.directory(),
  get.one.set.of.parameters = F,
  throw.error.if.incomplete = T
)

Description

'Transmutation' of a simulation is calculating a set of outcomes for another version without re-running the (expensive) differential equation solver

Usage

can.transmute.versions(
  from.version,
  to.version,
  from.sub.version = NULL,
  to.sub.version = NULL,
  print.details = T
)

Arguments

Description

Clear JHEEM Version, Calibration, Intervention, and Ontology Mappings

Usage

clear.all.managers()

Description

Clear a Calibration Cache

Usage

clear.calibration.cache(
  version,
  location,
  calibration.code,
  root.dir = get.jheem.root.directory("Cannot set up calibration: "),
  remove.mcmc.summary = F,
  allow.remove.incomplete = F
)

Arguments

Description

Clear JHEEM Calibrations

Usage

clear.calibrations()

Description

Clear JHEEM Interventions

Usage

clear.interventions()

Description

Clear Ontology Error Debug Info

Usage

clear.ontology.error.debug.info()

Description

Clear All Registered Ontology Mappings

Usage

clear.ontology.mappings()

Details

Clears all registered ontology mappings, such that new mappings can be added

Description

Clear a Calibration Cache

Usage

clear.transmute.calibration.cache(
  transmute.code,
  location,
  n.sim,
  to.version = NULL,
  to.sub.version = NULL,
  allow.remove.incomplete = F,
  root.dir = get.jheem.root.directory()
)

Arguments

Description

Clear JHEEM Versions

Usage

clear.versions()

Description

Collapse an Array According to Indices

Usage

collapse.array.according.to.indices(
  arr,
  small.indices,
  large.indices,
  small.n,
  check.consistency = T
)

Description

Combine multiple ontology mappings into a single mapping

Usage

combine.ontology.mappings(..., thorough.validation = T)

Arguments

Description

Compare how two simsets score on a likelihood.

Usage

compare.sims(likelihood, sim1, sim2, piecewise = T, log = F)

Arguments

Description

Compute Likelihood

Usage

compute.likelihood(
  likelihood,
  sim,
  log = T,
  use.optimized.get = F,
  check.consistency = T,
  debug = F
)

Arguments

Description

Compute Joint Likelihood Piecewise

Usage

compute.likelihood.piecewise(
  likelihood,
  sim,
  log = T,
  use.optimized.get = F,
  check.consistency = T,
  debug = F
)

Arguments

Details

For likelihoods that are not joint likelihoods, this is the same as compute.likelihood()

Description

Copy Calibration Info to a New Code

Usage

copy.calibration.info(
  from.code,
  to.code,
  n.iter = NULL,
  n.burn = NULL,
  thin = NULL,
  solver.metadata = NULL
)

Arguments

Description

Create a copy of a JHEEM Data Manager

Usage

copy.data.manager(
  data.manager = get.default.data.manager(),
  name,
  description,
  set.as.default = F,
  preserve.original.creation.date = T,
  preserve.original.last.modified.date = T
)

Arguments

Description

Make a copy of a Simulation Set

Usage

copy.simulation.set(simset, update.labels = T)

Arguments

Description

Create JHEEM Basic Likelihood Instructions

Usage

create.basic.likelihood.instructions(
  outcome.for.data,
  outcome.for.sim,
  dimensions = character(0),
  denominator.dimensions = dimensions,
  dimension.values = NULL,
  levels.of.stratification = NULL,
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  minimum.error.sd = 0,
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  error.variance.term = NULL,
  error.variance.type = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  name = outcome.for.sim,
  na.rm = NULL
)

Arguments

Details

Error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with an 'error.variance.type' of "cv" and an 'error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like an 'error.variance.type' of 'c("data.sd", "function.sd") and an 'error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

See Also

Other Basic Likelihoods: create.basic.likelihood.instructions.with.included.multiplier(), create.basic.likelihood.instructions.with.specified.outcome(), create.time.lagged.comparison.likelihood.instructions(), create.time.lagged.comparison.likelihood.instructions.with.included.multiplier()

Description

Create JHEEM Basic Likelihood Instructions With Included Multiplier

Usage

create.basic.likelihood.instructions.with.included.multiplier(
  outcome.for.data,
  outcome.for.sim,
  dimensions = character(0),
  denominator.dimensions = dimensions,
  dimension.values = NULL,
  levels.of.stratification = NULL,
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  minimum.error.sd = 0,
  included.multiplier,
  included.multiplier.sd,
  included.multiplier.correlation = NULL,
  included.multiplier.correlation.structure = c("compound.symmetry",
    "autoregressive.1")[1],
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  error.variance.term = NULL,
  error.variance.type = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  name = outcome.for.sim,
  na.rm = NULL
)

Arguments

Details

Error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with an 'error.variance.type' of "cv" and an 'error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like an 'error.variance.type' of 'c("data.sd", "function.sd") and an 'error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

See Also

Other Basic Likelihoods: create.basic.likelihood.instructions(), create.basic.likelihood.instructions.with.specified.outcome(), create.time.lagged.comparison.likelihood.instructions(), create.time.lagged.comparison.likelihood.instructions.with.included.multiplier()

Description

Create JHEEM Basic Likelihood Instructions With Specified Outcome

Usage

create.basic.likelihood.instructions.with.specified.outcome(
  outcome.for.sim,
  outcome.value,
  from.year,
  to.year,
  omit.years = NULL,
  minimum.error.sd = 0,
  correlation.different.years = 0.5,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  error.variance.term = NULL,
  error.variance.type = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  name = outcome.for.sim,
  na.rm = NULL
)

Arguments

Details

Error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with an 'error.variance.type' of "cv" and an 'error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like an 'error.variance.type' of 'c("data.sd", "function.sd") and an 'error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

See Also

Other Basic Likelihoods: create.basic.likelihood.instructions(), create.basic.likelihood.instructions.with.included.multiplier(), create.time.lagged.comparison.likelihood.instructions(), create.time.lagged.comparison.likelihood.instructions.with.included.multiplier()

Description

Create Basic Ratio Likelihood Instructions

Usage

create.basic.ratio.likelihood.instructions(
  outcome.for.data,
  outcome.for.sim,
  dimensions = character(0),
  denominator.dimensions = dimensions,
  dimension.values = NULL,
  levels.of.stratification = NULL,
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  error.variance.term = NULL,
  error.variance.type = NULL,
  ratio.cv = NULL,
  ratio.correlation = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  name = outcome.for.sim
)

Arguments

Description

Create Bernoulli Likelihood Instructions

Usage

create.bernoulli.likelihood.instructions(
  outcome.for.sim,
  years,
  probability.decreasing,
  weights,
  dimensions = character(0),
  levels.of.stratification = NULL,
  name = outcome.for.sim
)

Arguments

Description

Create instructions for a likelihood that computes using a custom function.

Usage

create.custom.likelihood.instructions(
  name,
  compute.function,
  get.data.function = NULL,
  weights = NULL,
  verbose = F
)

Arguments

Details

All weights used in the compute function should arrive in its "weights" argument in two ways: From the "weights" argument of the likelihood instructions (which apply to these instructions only), and additional weights passed top-down from joint likelihoods that will include this likelihood. So, no weights should be passed in the "data" function, for example, to maintain consistency and accuracy.

Description

A map from names to data managers. Pass it to using.data.binding or set.data.binding to build model objects with it.

Usage

create.data.binding(...)

Arguments

Value

An object of class 'jheem.data.binding'

Description

Create a JHEEM Data Manager

Usage

create.data.manager(name, description, set.as.default = F)

Arguments

Description

Turns 'variable' into an active binding whose value is get.data.manager(name), so code that reads the variable (for example CENSUS.MANAGER) gets the manager bound to 'name' in the active binding, and the read is recorded. Assigning a data manager to the variable registers it under 'name' (see register.data.manager); this is not allowed while a binding other than the session's is active or jheem2 is building under one.

If the variable already holds a data manager, it is registered under 'name' unless another manager already is. Aliasing a variable again to the same name does nothing; aliasing it to a different name is an error.

Usage

create.data.manager.alias(variable, name, envir = globalenv())

Arguments

Value

NULL, invisibly

Description

Create instructions that can try a sequence of likelihood instructions until the first one that works without error.

Usage

create.ifelse.likelihood.instructions(..., name = NULL)

Arguments

Details

The instructions are attempted in the order that they are supplied.

Description

Create an Intervention

Usage

create.intervention(
  ...,
  code = NULL,
  name = NULL,
  parameters = NULL,
  parameter.distribution = NULL,
  generate.parameters.function = NULL,
  overwrite.existing.intervention = F
)

Arguments

Details

Creates a 'jheem.intervention' object where all the intervention effects given in ... apply to all the target populations given in ...

Description

Create an Effect for an Intervention to Have on a Model Quantity

Usage

create.intervention.effect(
  quantity.name,
  start.time,
  effect.values,
  times,
  end.time = Inf,
  apply.effects.as,
  scale,
  allow.values.less.than.otherwise,
  allow.values.greater.than.otherwise,
  bindings = NULL,
  ...
)

Arguments

Description

Create an Engine to Run JHEEM Simulations

Usage

create.jheem.engine(
  version,
  location,
  end.year,
  sub.version = NULL,
  max.run.time.seconds = Inf,
  solver.metadata = NULL
)

Arguments

Description

Create a Model Specification for Running the JHEEM

Usage

create.jheem.specification(
  version,
  iteration,
  description,
  start.year,
  parent.version = NULL,
  do.not.inherit.model.quantity.names = character(),
  do.not.inherit.model.outcome.names = character(),
  do.not.inherit.transitions.for.dimension = character(),
  do.not.inherit.components.with.tags = character(),
  compartments.for.infected.only = list(),
  compartments.for.uninfected.only = list(),
  compartments.for.infected.and.uninfected = list(),
  compartments.for.outcomes = list(),
  age.endpoints = NULL,
  compartment.value.aliases = list(),
  order.locations.by.outcomes = NULL,
  labels = NULL,
  default.solver.metadata = NULL
)

Arguments

Description

Create a Transmuter

Usage

create.jheem.transmuter(
  simulation.set,
  to.version,
  from.year = simulation.set$from.year,
  to.year = simulation.set$to.year,
  to.sub.version = NULL
)

Arguments

Details

A jheem.transmuter can convert simulations from one version/specification to another using its transmute() method

Description

Create Weights to Be Applied in Calculating a JHEEM Likelihood

Usage

create.likelihood.weights(
  total.weight = 1,
  dimension.values = list(),
  is.recursive = T
)

Arguments

Description

Create an Functional Form that is Linear (possibly on a transformed scale)

Usage

create.linear.functional.form(
  intercept,
  slope,
  anchor.year,
  link = c("identity", "log", "logit")[1],
  min = NA,
  max = NA,
  intercept.link = NULL,
  intercept.min = NA,
  intercept.max = NA,
  intercept.robust.to.negative = F,
  slope.link = intercept.link,
  slope.min = NA,
  slope.max = NA,
  slope.robust.to.negative = F,
  parameters.are.on.transformed.scale = T,
  overwrite.parameters.with.alphas = F,
  error.prefix = "Cannot create linear functional.form: "
)

Arguments

See Also

create.static.functional.form, create.log.linear.functional.form, create.logistic.linear.functional.form

Description

Create a Linear Spline Functional Form

Usage

create.linear.spline.functional.form(
  knot.times,
  knot.values,
  link = c("identity", "log", "logit")[1],
  min = NA,
  max = NA,
  knot.link = link,
  knot.min = min,
  knot.max = max,
  knots.are.on.transformed.scale = F,
  overwrite.knot.values.with.alphas = F,
  before.time = NULL,
  before.modifier = NULL,
  after.time = NULL,
  after.modifier = NULL,
  modifiers.apply.to.change = T,
  modifier.link = link,
  after.modifier.increasing.change.link = modifier.link,
  after.modifier.decreasing.change.link = modifier.link,
  before.modifier.increasing.change.link = modifier.link,
  before.modifier.decreasing.change.link = modifier.link,
  modifier.min = min,
  modifier.max = max,
  overwrite.modifiers.with.alphas = F,
  error.prefix = "Cannot create linear spline functional.form: "
)

Arguments

Description

Create instructions that can instantiate a different 'jheem.likelihood' depending on the location.

Usage

create.location.based.ifelse.likelihood.instructions(
  ...,
  name = NULL,
  locations.list = list()
)

Arguments

Description

Create a Functional Form that is Linear on the Log Scale

Usage

create.log.linear.functional.form(
  intercept,
  slope,
  anchor.year,
  min = 0,
  max = Inf,
  parameters.are.on.log.scale = T,
  overwrite.parameters.with.alphas = F,
  error.prefix = "Cannot create log-linear functional.form: "
)

Arguments

See Also

create.static.functional.form, create.linear.functional.form, create.logistic.linear.functional.form

Description

Create a Functional Form that is Linear on the Logistic Scale

Usage

create.logistic.linear.functional.form(
  intercept,
  slope,
  anchor.year,
  min = 0,
  max = 1,
  parameters.are.on.logit.scale = T,
  overwrite.parameters.with.alphas = F,
  error.prefix = "Cannot create logistic-linear functional.form: "
)

Arguments

See Also

create.static.functional.form, create.linear.functional.form, create.log.linear.functional.form

Description

Create a Logistic Spline Functional Form

Usage

create.logistic.spline.functional.form(
  knot.times,
  knot.values,
  min = -Inf,
  max = Inf,
  knot.link = c("identity", "log", "logit")[2],
  knot.min = NA,
  knot.max = NA,
  knots.are.on.transformed.scale = F,
  overwrite.knot.values.with.alphas = F,
  fraction.of.asymptote.after.end = 0.05,
  fraction.of.asymptote.before.start = 0.025,
  fraction.of.asymptote.for.change.dir = 0.02,
  fraction.asymptote.link = "log",
  overwrite.fraction.asymptote.with.alphas = F,
  error.prefix = "Cannot create logistic spline functional.form: "
)

Arguments

Description

Create a "Logistic Tail" Functional Form

Usage

create.logistic.tail.functional.form(
  intercept,
  slope,
  anchor.year,
  logistic.after.frac.of.span = 0.5,
  min = 0,
  max = 1,
  intercept.link = "logistic",
  intercept.min = NA,
  intercept.max = NA,
  intercept.robust.to.negative = F,
  slope.link = intercept.link,
  slope.min = NA,
  slope.max = NA,
  slope.robust.to.negative = F,
  parameters.are.on.transformed.scale = F,
  overwrite.parameters.with.alphas = F,
  error.prefix = "Cannot create logistic-tail functional.form: "
)

Arguments

Details

A "logistic tail" functional form follows a linear (intercept + slope) in its first part, and a logistic-linear form in its second part. This allows a linear trend to taper off

Description

Create a Model Foreground Object

Usage

create.model.foreground(..., error.prefix = "")

Arguments

Details

A model foreground object applies multiple effects to multiple subpopulations *simultaneously*, with all effects targeting the same model quantity

Description

Create a "Guess-and-Check" Intervention that Must Satisfy Some Criteria by Varying Parameters

Usage

create.monotonic.criteria.based.intervention(
  base.intervention,
  completion.criteria,
  max.iterations,
  n.iterations.after.satisfying.criteria = 5,
  max.iterations.first.sim = max.iterations,
  n.iterations.after.satisfying.criteria.first.sim = ceiling(max.iterations.first.sim/2),
  max.failure.rate = 0,
  discount.prior.n = 0.5,
  code = NULL,
  name = NULL,
  overwrite.existing.intervention = F
)

Arguments

Description

Create a criterion for monotonic parameter, criteria-based interventions

Usage

create.monotonic.criterion(
  parameter.name,
  outcome,
  parameter.scale,
  parameter.initial.value,
  min.acceptable.parameter.value = NULL,
  max.acceptable.parameter.value = NULL,
  target.value,
  min.acceptable.value,
  max.acceptable.value,
  subsequent.thresholds.apply.after.iteration = numeric(),
  dimension.values = list(),
  draw.parameter.from.previous.sims = !is.function(parameter.initial.value),
  ...
)

Arguments

Details

If the intervention algorithm finds it needs to keep raising (or lowering) the parameter value so high (or low) that it reaches the bounds, the parameter will be considered optimized. This often occurs when the parameter's outcome is already beyond the target and would be generally deemed acceptable anyway.

Description

Create a Natural Spline Functional Form

Usage

create.natural.spline.functional.form(
  knot.times,
  knot.values,
  link = c("identity", "log", "logit")[1],
  min = NA,
  max = NA,
  knot.link = link,
  knot.min = min,
  knot.max = max,
  knots.are.on.transformed.scale = F,
  overwrite.knot.values.with.alphas = F,
  before.time = NULL,
  before.modifier = NULL,
  after.time = NULL,
  after.modifier = NULL,
  modifiers.apply.to.change = T,
  modifier.link = link,
  after.modifier.increasing.change.link = modifier.link,
  after.modifier.decreasing.change.link = modifier.link,
  before.modifier.increasing.change.link = modifier.link,
  before.modifier.decreasing.change.link = modifier.link,
  modifier.min = min,
  modifier.max = max,
  overwrite.modifiers.with.alphas = F,
  error.prefix = "Cannot create natural spline functional.form: "
)

Arguments

Description

Create Nested Proportion Likelihood Instructions

Usage

create.nested.proportion.likelihood.instructions(
  outcome.for.data,
  denominator.outcome.for.data,
  outcome.for.sim,
  outcome.for.n.multipliers = denominator.outcome.for.data,
  location.types,
  minimum.geographic.resolution.type,
  maximum.locations.per.type = 3,
  dimensions = NULL,
  levels.of.stratification = 0:length(dimensions),
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  exclude.denominator.ontology.names = NULL,
  location.overall.keep.threshold = 5,
  location.stratum.keep.threshold = 1,
  p.bias.inside.location,
  p.bias.outside.location,
  p.bias.sd.inside.location,
  p.bias.sd.outside.location,
  within.location.p.error.correlation = 0.5,
  within.location.n.error.correlation = 0.5,
  minimum.error.sd = 0,
  correlation.different.locations = 0,
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  obs.n.correlation.different.years = within.location.n.error.correlation,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  n.multiplier.cv = 0.1,
  p.error.variance.term = NULL,
  p.error.variance.type = NULL,
  n.error.variance.term = 0.05,
  n.error.variance.type = "cv",
  n.error.variance.term.if.estimated = 0.8,
  n.error.variance.type.if.estimated = "exp.of.variance",
  weights = list(),
  equalize.weight.by.year = T,
  partitioning.function,
  name = outcome.for.sim
)

Arguments

Details

P error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with a 'p.error.variance.type' of "cv" and a 'p.error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like a 'p.error.variance.type' of 'c("data.sd", "function.sd") and a 'p.error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

N error variance works the same way but only one type is currently supported: "cv" (covariance).

See Also

Other Nested Proportion Likelihoods: create.nested.proportion.likelihood.instructions.with.included.multiplier(), create.time.lagged.comparison.nested.proportion.likelihood.instructions(), get.p.bias.estimates()

Description

Create Nested Proportion Likelihood Instructions With Included Multiplier

Usage

create.nested.proportion.likelihood.instructions.with.included.multiplier(
  outcome.for.data,
  denominator.outcome.for.data,
  outcome.for.sim,
  outcome.for.n.multipliers = denominator.outcome.for.data,
  location.types,
  minimum.geographic.resolution.type,
  maximum.locations.per.type = 3,
  dimensions = NULL,
  levels.of.stratification = 0:length(dimensions),
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  exclude.denominator.ontology.names = NULL,
  location.overall.keep.threshold = 5,
  location.stratum.keep.threshold = 1,
  included.multiplier,
  included.multiplier.sd,
  included.multiplier.correlation = NULL,
  included.multiplier.correlation.structure = c("compound.symmetry",
    "autoregressive.1")[1],
  p.bias.inside.location,
  p.bias.outside.location,
  p.bias.sd.inside.location,
  p.bias.sd.outside.location,
  within.location.p.error.correlation = 0.5,
  within.location.n.error.correlation = 0.5,
  minimum.error.sd = 0,
  correlation.different.locations = 0,
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  obs.n.correlation.different.years = within.location.n.error.correlation,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  n.multiplier.cv = 0.1,
  p.error.variance.term = NULL,
  p.error.variance.type = NULL,
  n.error.variance.term = 0.05,
  n.error.variance.type = "cv",
  n.error.variance.term.if.estimated = 0.8,
  n.error.variance.type.if.estimated = "exp.of.variance",
  weights = list(),
  equalize.weight.by.year = T,
  partitioning.function,
  name = outcome.for.sim
)

Arguments

Details

P error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with a 'p.error.variance.type' of "cv" and a 'p.error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like a 'p.error.variance.type' of 'c("data.sd", "function.sd") and a 'p.error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

N error variance works the same way but only one type is currently supported: "cv" (covariance).

See Also

Other Nested Proportion Likelihoods: create.nested.proportion.likelihood.instructions(), create.time.lagged.comparison.nested.proportion.likelihood.instructions(), get.p.bias.estimates()

Description

Create a mapping that allows transformations of arrays

Usage

create.ontology.mapping(
  mappings,
  from.dimensions,
  to.dimensions = from.dimensions,
  name = NULL,
  error.prefix = "Error creating ontology mapping: "
)

Arguments

Description

Create an ontology mapping that lumps otherwise unspecified categories into an "other" category

Usage

create.other.catchall.ontology.mapping(
  dimension,
  from.values,
  to.values,
  other.category.value = "other"
)

Arguments

Description

Create Outcome Location Mapping

Usage

create.outcome.location.mapping(
  location.mappings,
  outcome.name,
  version = jheem.kernel$version,
  location = jheem.kernel$location,
  sub.version = NULL,
  jheem.kernel = NULL
)

Arguments

Usage

create.outcome.metadata(
  display.name,
  description,
  scale,
  axis.name,
  units,
  singular.unit = units,
  display.as.percent = scale == "proportion"
)

Arguments

An object of class 'outcome.metadata' which can be passed to track.transition, track.quantity, track.dynamic.quantity

Create a metadata object for tracked quantities or tracked transitions

Description

Create a Collection of JHEEM Simulation-Sets

Usage

create.simset.collection(
  version,
  calibration.code,
  locations,
  interventions,
  n.sim = NULL,
  sub.version = NULL,
  root.dir = get.jheem.root.directory("Cannot create simset collection: ")
)

Arguments

Value

An object of class 'jheem.simset.collection'

Description

Create Solver Metadata

Usage

create.solver.metadata(
  method = "ode45",
  package = NULL,
  atol = 0.01,
  rtol = 0.001
)

Description

Create an Static (Time-Invariant) Functional Form

Usage

create.static.functional.form(
  value,
  link = c("identity", "log", "logit")[1],
  min = NA,
  max = NA,
  value.name = "value",
  value.is.on.transformed.scale = F,
  overwrite.parameters.with.alphas = F,
  error.prefix = "Cannot create static functional.form: "
)

Arguments

Description

Create A Style Manager

Usage

create.style.manager(
  color.sim.by = "stratum",
  color.data.by = "stratum",
  shade.data.by = "location",
  linetype.sim.by = "simset",
  shape.data.by = "source",
  shape.sim.by = "simset",
  sim.palette = ggsci::pal_jama(),
  data.palette = ggsci::pal_jama(),
  linetypes = c("solid", "dashed", "dotted", "dotdash"),
  shapes = 21:25,
  shade.increment = -64,
  alpha.ribbon = 0.2,
  alpha.line = 1,
  linewidth.baseline = 1,
  linewidth.slope = 3
)

Arguments

Description

Create a Target Population to Which Interventions May be Applied

Usage

create.target.population(..., name, invert = F)

Arguments

See Also

Other Functions to create Target Populations: diff.target.populations(), intersect.target.populations(), union.target.populations()

Description

Create JHEEM Basic Likelihood Instructions

Usage

create.time.lagged.comparison.likelihood.instructions(
  outcome.for.data,
  outcome.for.sim,
  dimensions = character(0),
  denominator.dimensions = dimensions,
  dimension.values = NULL,
  levels.of.stratification = NULL,
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  minimum.error.sd = 0,
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  error.variance.term = NULL,
  error.variance.type = NULL,
  ratio.cv = NULL,
  ratio.correlation = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  use.lognormal.approximation = T,
  name = outcome.for.sim,
  na.rm = NULL
)

Arguments

Details

Error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with an 'error.variance.type' of "cv" and an 'error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like an 'error.variance.type' of 'c("data.sd", "function.sd") and an 'error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

See Also

Other Basic Likelihoods: create.basic.likelihood.instructions(), create.basic.likelihood.instructions.with.included.multiplier(), create.basic.likelihood.instructions.with.specified.outcome(), create.time.lagged.comparison.likelihood.instructions.with.included.multiplier()

Description

Create JHEEM Basic Likelihood Instructions

Usage

create.time.lagged.comparison.likelihood.instructions.with.included.multiplier(
  outcome.for.data,
  outcome.for.sim,
  dimensions = character(0),
  denominator.dimensions = dimensions,
  dimension.values = NULL,
  levels.of.stratification = NULL,
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  minimum.error.sd = 0,
  included.multiplier,
  included.multiplier.sd,
  included.multiplier.correlation = NULL,
  included.multiplier.correlation.structure = c("compound.symmetry",
    "autoregressive.1")[1],
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  error.variance.term = NULL,
  error.variance.type = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  use.lognormal.approximation = T,
  name = outcome.for.sim,
  na.rm = NULL
)

Arguments

Details

Error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with an 'error.variance.type' of "cv" and an 'error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like an 'error.variance.type' of 'c("data.sd", "function.sd") and an 'error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

See Also

Other Basic Likelihoods: create.basic.likelihood.instructions(), create.basic.likelihood.instructions.with.included.multiplier(), create.basic.likelihood.instructions.with.specified.outcome(), create.time.lagged.comparison.likelihood.instructions()

Description

Create Time Lagged Comparison Nested Proportion Likelihood Instructions

Usage

create.time.lagged.comparison.nested.proportion.likelihood.instructions(
  outcome.for.data,
  denominator.outcome.for.data,
  outcome.for.sim,
  outcome.for.n.multipliers = denominator.outcome.for.data,
  location.types,
  minimum.geographic.resolution.type,
  maximum.locations.per.type = 3,
  dimensions = NULL,
  levels.of.stratification = 0:length(dimensions),
  from.year = -Inf,
  to.year = Inf,
  omit.years = NULL,
  sources.to.use = NULL,
  exclude.denominator.ontology.names = NULL,
  location.overall.keep.threshold = 5,
  location.stratum.keep.threshold = 1,
  p.bias.inside.location,
  p.bias.outside.location,
  p.bias.sd.inside.location,
  p.bias.sd.outside.location,
  within.location.p.error.correlation = 0.5,
  within.location.n.error.correlation = 0.5,
  minimum.error.sd = 0,
  correlation.different.locations = 0,
  correlation.different.years = 0.5,
  correlation.different.strata = 0.1,
  correlation.different.sources = 0.3,
  correlation.same.source.different.details = 0.3,
  obs.n.correlation.different.years = within.location.n.error.correlation,
  observation.correlation.form = c("compound.symmetry", "autoregressive.1")[1],
  n.multiplier.cv = 0.1,
  p.error.variance.term = NULL,
  p.error.variance.type = NULL,
  n.error.variance.term = 0.05,
  n.error.variance.type = "cv",
  n.error.variance.term.if.estimated = 0.8,
  n.error.variance.type.if.estimated = "exp.of.variance",
  ratio.cv = NULL,
  ratio.correlation = NULL,
  weights = list(),
  equalize.weight.by.year = T,
  partitioning.function,
  use.lognormal.approximation = T,
  name = outcome.for.sim
)

Arguments

Details

P error variance may be specified in a variety of ways, and more than one can be used in the same likelihood. 'error.variance.type' is a vector containing single character values describing the type(s) of variance to use, and there is one 'error.variance.term' corresponding to each. The types "sd" (standard deviation), "variance", "cv" (covariance), and "exp.of.variance" (exponent of variance) require a single numeric value as their term. The types "data.sd", "data.variance", and "data.cv" indicate that error data will be pulled from the data manager and matched directly to data points, and they require NULL as their term. Finally, the type "function.sd" requires as its term a function that takes only arguments "data", "details", "version", and "location" and returns a numeric array of the same dimensions as "data". Supplying multiple types (and therefore terms) results in the variances created by each being summed prior to use in the measurement error covariance matrix.

As an example, one type could be specified with a 'p.error.variance.type' of "cv" and a 'p.error.variance.term' of 0.13 (or list(0.13)). Two or more types might look like a 'p.error.variance.type' of 'c("data.sd", "function.sd") and a 'p.error.variance.term' of list(NULL, my_sd_function). It is possible to use the same type multiple times, although this is not recommended.

N error variance works the same way but only one type is currently supported: "cv" (covariance).

See Also

Other Nested Proportion Likelihoods: create.nested.proportion.likelihood.instructions(), create.nested.proportion.likelihood.instructions.with.included.multiplier(), get.p.bias.estimates()

Description

While a kernel for this specification's version is built under a binding, the binding must contain every declared name, and resolving any other name is an error. With no binding active, the declaration is not checked.

Usage

declare.data.managers(specification, names)

Arguments

Value

NULL, invisibly

Description

Create a Target Population that is the DIFFERENCE between two other Target Populations

Usage

## S3 method for class 'target.populations'
diff(pop1, pop2, name = NULL)

Arguments

Details

The resulting target population contains everyone who IS in pop1 but NOT in pop2

See Also

Other Functions to create Target Populations: create.target.population(), intersect.target.populations(), union.target.populations()

Description

Disable JHEEM Solver Tracking

Usage

disable.jheem.solver.tracking()

Description

This function redistributes certain values in a dimension across the remaining values in that dimension. This is useful to remove problematic dimension values – like race = 'unknown' – while keeping the marginal sums and the relevant ratios between the remaining dimension values the same as they were before.

Usage

distribute.dimension.values(
  arr,
  dimension.values.to.distribute,
  round.distributed.dimension.values
)

Arguments

Details

If multiple dimensions should be redistributed, they will be redistributed in the order that they exist in 'dimension.values.to.distribute'. This can result in different outputs for different orders.

Value

An array with the same number of dimensions as 'arr', but without the dimension values specified in 'dimension.values.to.distribute'.

Description

Enable JHEEM Solver Tracking

Usage

enable.jheem.solver.tracking()

Description

For use in figuring out why a call to get.mappings.to.align.ontologies did not work

If no mappings can be found, prints detailed information on why we were unable to find mappings

Usage

examine.get.mappings.to.align.ontologies(
  ontology.1,
  ontology.2,
  align.on.dimensions = intersect(names(ontology.1), names(ontology.2)),
  allow.non.overlapping.incomplete.dimensions = F,
  include.dim.names = NULL,
  print.ontologies = T
)

Arguments

Description

For use in figuring out why a call to get.ontology.mapping did not work

If no mapping can be found, prints detailed information on why we were unable to find a mapping

Usage

examine.get.ontology.mapping(
  from.ontology,
  to.ontology,
  allow.non.overlapping.incomplete.dimensions = F,
  print.ontologies = T
)

Arguments

Description

Prints detailed information on why we were unable to find a mapping

Usage

examine.recent.failed.ontology.mapping(which = 1, print.ontologies = T)

Arguments

Description

Execute Plotly Plot A utility that performs the plot-rendering half of the plot simulations in plotly operation for applications like webtools that need the step to be separate.

Usage

execute.plotly.plot(
  prepared.plot.data,
  outcomes = NULL,
  split.by = NULL,
  facet.by = NULL,
  plot.which = "sim.and.data",
  summary.type = c("individual.simulation", "mean.and.interval",
    "median.and.interval")[1],
  plot.year.lag.ratio = F,
  n.facet.rows = NULL,
  style.manager = get.default.style.manager(),
  debug = F,
  hide.legend = FALSE
)

Description

Execute Simplot A utility that performs the plot-rendering half of the simplot operation for applications like webtools that need the step to be separate.

Usage

execute.simplot(
  prepared.plot.data,
  outcomes = NULL,
  split.by = NULL,
  facet.by = NULL,
  plot.which = c("sim.and.data", "sim.only", "data.only")[1],
  summary.type = c("individual.simulation", "mean.and.interval",
    "median.and.interval")[1],
  plot.year.lag.ratio = F,
  n.facet.rows = NULL,
  style.manager = get.default.style.manager(),
  debug = F
)

Arguments

Description

Expand an Array to Greater Dimensions

Usage

expand.array(to.expand, target.dim.names)

Arguments

Description

Extract the Last Simulation From A Calibration

Usage

extract.last.simulation.from.calibration(
  version,
  location,
  calibration.code,
  root.dir = get.jheem.root.directory("Cannot set up calibration: "),
  allow.incomplete = T,
  include.first.sim = F,
  chains = NULL
)

Arguments

Description

Find Ontology Mapping

Usage

find.ontology.mapping(
  from.ontology,
  to.ontology,
  required.dimensions,
  required.dim.names,
  get.two.way.alignment,
  allow.non.overlapping.incomplete.dimensions,
  from.ontology.name = ifelse(get.two.way.alignment, "ontology.1", "the from.ontology"),
  to.ontology.name = ifelse(get.two.way.alignment, "ontology.2", "the to.ontology"),
  get.details = F
)

Arguments

Value

If get.two.way.alignment is false, returns a list of one or more ontology.mapping objects, such that applying these mappings to an array with from.dim.names ('from.arr') would yield an array with dimnames ('to.dim.names')

If get.two.way.alignment is true, returns a list with two elements, each of which is itself a list of one or more ontology.mapping objects, such that applying the first set of mappings to an array with from.dimnames and applying the second set of mappings to an array with to.dimnames would both yield arrays with dimensions 'align.on.dimensions', both with the same dimnames

Description

Summarize the progress of a calibration task

Usage

get.calibration.progress(
  version,
  locations,
  calibration.code,
  root.dir = get.jheem.root.directory(),
  round.to.digits = ifelse(as.pct, 0, 2),
  as.pct = T
)

Arguments

Description

Get a Set of Indices to Collapse One Array into Another

Usage

get.collapse.array.indices(large.arr.dim.names, small.arr.dim.names)

Arguments

Value

A list with four elements: $large.indices, $small.indices, $large.n, $small.n large.indices and small.indices are integer vectors of equal length, such that the value at large.indices[i] goes into the value at small.indices[i] if collapsing

Description

Get a function that produces a cubic spline over the *cumulative* counts by age up to a given age

Usage

get.cumulative.age.counts.smoother(
  counts,
  given.age.brackets,
  smooth.infinite.age.to = Inf,
  allow.extrapolation = F,
  method = c("monoH.FC", "hyman")[1],
  error.prefix = ""
)

Arguments

Description

Get the active data binding

Usage

get.data.binding()

Value

The binding made active most recently by using.data.binding or a jheem2 function building under it; otherwise the session binding if set.data.binding was called; otherwise NULL

Description

Get the identity of a data binding

Usage

get.data.binding.identity(binding)

Arguments

Value

A data frame with one row per name: name, data.manager.name, and the file, release, repository, sha256, md5, and modified status of the data manager bound to it (NA where unknown; see get.data.manager.provenance)

Description

The data managers that were in use when a kernel, likelihood, simulation set, or calibration was built: those registered by name, the default data manager, a manager passed in explicitly, and every manager loaded with load.data.manager in the session. The record can include managers the artifact did not use.

Usage

get.data.inputs(x)

Arguments

Value

NULL if nothing was recorded (for example, for artifacts built by earlier versions of jheem2). Otherwise a list with

coverage

'declared' if any managers were registered by name, 'loaded', 'bound' if the artifact was built under a data binding, and 'unknown' if any listed manager has no identity (it was not loaded from a file)

recorded.at

When the record was made

run.record

The JHEEM_RUN_RECORD environment variable at that time, if it was set: a pointer to the run's own record, for example a container run's records directory

data.managers

A data frame with one row per data manager: registered.as (the registered names, NA if none), default, passed, data.manager.name, file, release, repository, sha256, md5, modified (whether its contents had changed since loading; NA when not known) and loaded.at; and, for an artifact built under a data binding, bound.as (the names it was bound under, NA if none) and used (whether the build resolved any of those names; NA if not bound)

Description

With a data binding active (see get.data.binding), the data manager bound to the name, and the name is recorded as used by any model object being built. Otherwise, the data manager registered under the name.

Usage

get.data.manager(name)

Arguments

Value

A jheem.data.manager

Description

Where a data manager came from: the file it was loaded from, an MD5 digest of that file, the release it belongs to when the loader supplied it, when it was loaded, and whether its contents have changed since.

Usage

get.data.manager.provenance(data.manager)

Arguments

Value

A list with elements file, md5, sha256, release, repository, loaded.at and modified, or NULL for a data manager that was not loaded from a file (or was loaded by an earlier version of jheem2)

Description

Get pretty names for outcomes

Usage

get.data.outcome.pretty.names(
  data.manager = get.default.data.manager(),
  outcomes
)

get.data.outcome.labels(data.manager = get.default.data.manager(), outcomes)

get.data.outcome.descriptions(
  data.manager = get.default.data.manager(),
  outcomes
)

Arguments

Details

Gets the pretty.names, labels, or descriptions registered for the outcomes with the data manager

Value

A character vector of pretty names, labels, or descriptions

Functions

• get.data.outcome.labels(): Get the labels for outcomes

• get.data.outcome.descriptions(): Get descriptions for outcomes

Description

Get full names for data sources

Usage

get.data.source.full.names(data.manager = get.default.data.manager(), sources)

get.data.source.short.names(data.manager = get.default.data.manager(), sources)

Arguments

Details

Gets the full.names or short.names for the data sources in the data manager

Value

A character vector of full names or short names

Functions

• get.data.source.short.names(): Get short names for data sources

Description

Get the default data manager

Usage

get.default.data.manager()

Description

Get the Default Style Manager

Usage

get.default.style.manager(type = c("ggplot", "plotly")[1])

Description

Get the Array Indices Under Expanded Dim-names

Usage

get.expand.array.indices(to.expand.dim.names, target.dim.names, index.from = 1)

Description

Get a Set of Indices to Collapse One Array into Another, expanding the first array if needed to add dimensions

Usage

get.expand.then.collapse.array.indices(
  large.arr.dim.names,
  small.arr.dim.names
)

Arguments

Value

A list with four elements: $large.indices, $small.indices, $large.n, $small.n large.indices and small.indices are integer vectors of equal length, such that the value at large.indices[i] goes into the value at small.indices[i] if collapsing

Description

Get an "Identity" Ontology Mapping (that does not change data structure when applied)

Usage

get.identity.ontology.mapping()

Value

An object of class "ontology.mapping"

See Also

Other Functions to identify why ontologies fail to map: apply.ontology.mapping(), map.value.ontology()

Description

Get JHEEM Root Directory #' @description Gets the directory that calibration directories and files will be written into, if it has been set with "set.jheem.root.directory"

Usage

get.jheem.root.directory(error.prefix = "")

Description

Get JHEEM Solver Tracked Info

Usage

get.jheem.solver.tracked.info()

Description

Get locations with data for an outcome

Usage

get.locations.with.data(
  data.manager = get.default.data.manager(),
  outcome,
  metric = "estimate",
  years,
  exclude.ontology.names = NULL
)

Arguments

Details

Get locations that have data for a given outcome, metric and set of years. Searches all sources and all ontologies that haven't been specifcally excluded

Description

Get a Pair of Ontology Mappings that Aligns two Data Elements

Usage

get.mappings.to.align.ontologies(
  ontology.1,
  ontology.2,
  align.on.dimensions = intersect(names(ontology.1), names(ontology.2)),
  allow.non.overlapping.incomplete.dimensions = F,
  include.dim.names = NULL
)

Arguments

Value

Either a (1) NULL - if no mappings could align the ontologies or (2) a list with two elements, each an ontology.mapping object, such that applying the first to data conforming to ontology.1 and applying the second to data conforming to ontology.2 yields two data objects with the same dim.names

Description

For one chain, the mixing statistic is the variance in the first quarter of simulations divided by the variance in the last quarter of simulations.

Usage

get.mcmc.mixing.statistic(
  simset,
  match.names = NULL,
  chains = simset$unique.chains,
  sort = T
)

Arguments

Description

Returns information useful for debugging ontology mapping errors

Usage

get.ontology.error.debug.info()

Description

Get an Ontology Mapping to Transform Data

Usage

get.ontology.mapping(
  from.ontology,
  to.ontology,
  allow.non.overlapping.incomplete.dimensions = F
)

Arguments

Value

Either an object of class 'ontology.mapping' or NULL if no mapping that would bridge the differences is found

Description

Get a Matrix that Applies an Ontology Mapping

Usage

get.ontology.mapping.matrix(
  ontology.mapping,
  from.dim.names,
  to.dim.names = NULL,
  error.prefix = ""
)

Arguments

Description

Get P Bias Estimates

Usage

get.p.bias.estimates(
  data.manager = get.default.data.manager(),
  dimensions,
  levels.of.stratification,
  outcome.for.p,
  outcome.for.n,
  sub.location.type,
  super.location.type,
  main.location.type = "CBSA",
  minimum.sample.size = 12,
  from.year = NULL,
  to.year = NULL,
  main.location.type.p.source = NULL,
  sub.location.type.p.source = NULL,
  super.location.type.p.source = NULL,
  main.location.type.n.source = NULL,
  sub.location.type.n.source = NULL,
  super.location.type.n.source = NULL,
  debug = F
)

Arguments

Value

A list containing mean and sd for p-bias "inside" and p-bias "outside" estimates. Some or all of these may be NA if 'sub.location.type' and/or 'super.location.type' were left NULL.

See Also

Other Nested Proportion Likelihoods: create.nested.proportion.likelihood.instructions(), create.nested.proportion.likelihood.instructions.with.included.multiplier(), create.time.lagged.comparison.nested.proportion.likelihood.instructions()

Description

Get Prior JHEEM Model Versions

Usage

get.prior.versions(version, recursive = T)

See Also

Other JHEEM Version Management Functions: get.specification.for.version(), register.model.specification(), version.has.been.registered()

Description

param ontology.name The name of an ontology (must be registered in the data manager)

Usage

get.registered.ontology(
  data.manager = get.default.data.manager(),
  ontology.name
)

Description

Get Simulation Data

Usage

get.simset.data(
  simset,
  outcomes,
  output = c("value", "numerator", "denominator")[[1]],
  keep.dimensions = NULL,
  dimension.values = list(),
  ...,
  check.consistency = T,
  drop.single.outcome.dimension = T,
  error.prefix = "Error getting simulation results: "
)

Arguments

Details

Returns an array whose dimensions are keep.dimensions - plus (if there is more than one outcome or drop.single.outcome.dimensions==FALSE) an 'outcome' dimension at the end

Description

Get Simset Filename

Usage

get.simset.filename(
  version,
  location,
  calibration.code,
  n.sim,
  intervention.code = NULL,
  sub.version = NULL,
  include.path = T,
  root.dir = get.jheem.root.directory("Cannot get.simset.filename(): "),
  error.prefix = ""
)

Description

Get Simset Filename and Directory

Usage

get.simset.filename.and.dir(
  version,
  location,
  calibration.code,
  n.sim,
  intervention.code = NULL,
  sub.version = NULL,
  root.dir = get.jheem.root.directory("Cannot get.simset.filename(): "),
  error.prefix = ""
)

Description

Get Simulation Set Paramater Values

Usage

get.simset.params(
  simset,
  match.names = NULL,
  simulation.indices = simset$n.sim,
  drop = T
)

Arguments

Description

Get a Simulation Metadata Object

Usage

get.simulation.metadata(
  version,
  location,
  from.year = NULL,
  to.year = NULL,
  n.sim = 1,
  sub.version = NULL,
  jheem.kernel = NULL,
  update.labels = T,
  error.prefix = paste0("Error deriving the simulation-metadata for '", version,
    "' and location '", location, "': ")
)

Arguments

Details

A simulation.metadata object contains metadata, such as the dim.names to which its contents will conform

Description

Get a JHEEM Specification Object

Usage

get.specification.for.version(version)

Arguments

Value

An object of class 'jheem.specification'

See Also

Other JHEEM Version Management Functions: get.prior.versions(), register.model.specification(), version.has.been.registered()

Description

Get a Specification Metadata Object

Usage

get.specification.metadata(version, location, error.prefix = "")

Arguments

Details

A specification.metadata object contains metadata, such as dim.names of quantities, for a specification

Description

Get year bounds for outcome

Usage

get.year.bounds.for.outcome(
  data.manager = get.default.data.manager(),
  outcome,
  metric = "estimate",
  exclude.ontology.names = NULL
)

Arguments

Details

Gets the earliest and latest years for which a data manager has data for a given outcome and metric. Searches all sources and all ontologies that haven't been specifically excluded

Description

Import data from another JHEEM Data Manager

Usage

import.data(to.data.manager = get.default.data.manager(), from.data.manager)

Arguments

Description

Get an ontology's incomplete dimensions

Usage

incomplete.dimensions(x)

Arguments

Value

A vector containing the names of the ontology's incomplete dimensions

Description

Get an ontology's incomplete dimensions

Usage

## S3 method for class 'ontology'
incomplete.dimensions(x)

Arguments

Value

A vector containing the names of the ontology's incomplete dimensions

Description

Compares the percentage difference between the marginal sums of more stratified data and the data it corresponds to in more aggregated stratifications of the same source and ontology.

Usage

inspect.marginals(
  data.manager = get.default.data.manager,
  outcome = NULL,
  source = NULL,
  ontology = NULL,
  sub.stratification = NULL,
  super.stratification = NULL,
  years = NULL,
  locations = NULL,
  threshold.percent = 10,
  threshold.magnitude = 50,
  check.up.to.n.way.stratified = 1
)

Arguments

Description

Instantiate Likelihood

Usage

instantiate.likelihood(
  instructions,
  version,
  location,
  sub.version = NULL,
  data.manager = get.default.data.manager(),
  verbose = F,
  throw.error.if.no.data = F,
  error.prefix = "Error instantiating likelihood from instructions: "
)

Arguments

Description

Interpolate Values

Usage

interpolate(values, value.times, desired.times)

Description

Create a Target Population that is the INTERSECTION of Other Target Populations

Usage

intersect.target.populations(..., name = NULL)

Arguments