Package 'multimput' reference manual

Title:	Using Multiple Imputation to Address Missing Data
Description:	Accompanying package for the paper: Working with population totals in the presence of missing data comparing imputation methods in terms of bias and precision. Published in 2017 in the Journal of Ornithology volume 158 page 603–615 (<doi:10.1007/s10336-016-1404-9>).
Authors:	Thierry Onkelinx [aut, cre] (<https://orcid.org/0000-0001-8804-4216>, Research Institute for Nature and Forest (INBO)), Koen Devos [aut] (<https://orcid.org/0000-0001-7265-6349>, Research Institute for Nature and Forest (INBO)), Paul Quataert [aut] , Research Institute for Nature and Forest (INBO) [cph, fnd]
Maintainer:	Thierry Onkelinx <[email protected]>
License:	GPL-3
Version:	0.2.14
Built:	2024-10-04 12:23:46 UTC
Source:	https://github.com/inbo/multimput

Aggregate an imputed dataset

Description

Aggregate an imputed dataset

Usage

aggregate_impute(object, grouping, fun, filter = list(), join)

## S4 method for signature 'ANY'
aggregate_impute(object, grouping, fun, filter = list(), join)

## S4 method for signature 'rawImputed'
aggregate_impute(object, grouping, fun, filter = list(), join)

## S4 method for signature 'aggregatedImputed'
aggregate_impute(object, grouping, fun, filter = list(), join)
aggregate_impute(object, grouping, fun, filter = list(), join)

## S4 method for signature 'ANY'
aggregate_impute(object, grouping, fun, filter = list(), join)

## S4 method for signature 'rawImputed'
aggregate_impute(object, grouping, fun, filter = list(), join)

## S4 method for signature 'aggregatedImputed'
aggregate_impute(object, grouping, fun, filter = list(), join)

Arguments

`object`	A `rawImputed` object.
`grouping`	A vector of variables names to group the aggregation on.
`fun`	The function to aggregate.
`filter`	An optional argument to filter the raw dataset before aggregation. Will be passed to `dplyr::filter()`.
`join`	An optional argument to filter the raw dataset based on a data.frame. A `dplyr::semi_join()` will be applied with `join` or each element of `join` in case join is a list.

Examples

dataset <- generate_data(n_year = 10, n_site = 50, n_run = 1)
dataset$Count[sample(nrow(dataset), 50)] <- NA
model <- lm(Count ~ Year + factor(Period) + factor(Site), data = dataset)
imputed <- impute(data = dataset, model = model)
aggregate_impute(imputed, grouping = c("Year", "Period"), fun = sum)
dataset <- generate_data(n_year = 10, n_site = 50, n_run = 1)
dataset$Count[sample(nrow(dataset), 50)] <- NA
model <- lm(Count ~ Year + factor(Period) + factor(Site), data = dataset)
imputed <- impute(data = dataset, model = model)
aggregate_impute(imputed, grouping = c("Year", "Period"), fun = sum)

The `aggregatedImputed` class Holds an aggregated imputation data set

Description

The aggregatedImputed class Holds an aggregated imputation data set

Slots

Covariate: A data.frame with the covariates.
Imputation: A matrix with aggregated imputed values.

Generate data for a regular monitoring design. The counts follow a negative binomial distribution with given size parameters and the true mean mu depending on a year, period and site effect. All effects are independent from each other and have, on the log-scale, a normal distribution with zero mean and given standard deviation.

Usage

generate_data(
  intercept = 2,
  n_year = 24,
  n_period = 6,
  n_site = 20,
  year_factor = FALSE,
  period_factor = FALSE,
  site_factor = FALSE,
  trend = 0.01,
  sd_rw_year = 0.1,
  amplitude_period = 1,
  mean_phase_period = 0,
  sd_phase_period = 0.2,
  sd_site = 1,
  sd_rw_site = 0.02,
  sd_noise = 0.01,
  size = 2,
  n_run = 10,
  as_list = FALSE,
  details = FALSE
)
generate_data(
  intercept = 2,
  n_year = 24,
  n_period = 6,
  n_site = 20,
  year_factor = FALSE,
  period_factor = FALSE,
  site_factor = FALSE,
  trend = 0.01,
  sd_rw_year = 0.1,
  amplitude_period = 1,
  mean_phase_period = 0,
  sd_phase_period = 0.2,
  sd_site = 1,
  sd_rw_site = 0.02,
  sd_noise = 0.01,
  size = 2,
  n_run = 10,
  as_list = FALSE,
  details = FALSE
)

Arguments

`intercept`	The global mean on the log-scale.
`n_year`	The number of years.
`n_period`	The number of periods.
`n_site`	The number of sites.
`year_factor`	Convert year to a factor. Defaults to `FALSE`.
`period_factor`	Convert period to a factor. Defaults to `FALSE`.
`site_factor`	Convert site to a factor. Defaults to `FALSE`.
`trend`	The long-term linear trend on the log-scale.
`sd_rw_year`	The standard deviation of the year effects on the log-scale.
`amplitude_period`	The amplitude of the periodic effect on the log-scale.
`mean_phase_period`	The mean of the phase of the periodic effect among years. Defaults to `0`.
`sd_phase_period`	The standard deviation of the phase of the periodic effect among years.
`sd_site`	The standard deviation of the site effects on the log-scale.
`sd_rw_site`	The standard deviation of the random walk along year per site on the log-scale.
`sd_noise`	The standard deviation of the noise effects on the log-scale.
`size`	The size parameter of the negative binomial distribution.
`n_run`	The number of runs with the same mu.
`as_list`	Return the dataset as a list rather than a data.frame. Defaults to `FALSE`.
`details`	Add variables containing the year, period and site effects. Defaults tot `FALSE`.

Value

A data.frame with five variables. Year, Month and Site are factors identifying the location and time of monitoring. Mu is the true mean of the negative binomial distribution in the original scale. Count are the simulated counts.

Combine two models into a hurdle model

Description

Multiplies the imputed values for the presence model with those of the count model. Please make sure that the order of the observations in both models is identical. The resulting object will contain the union of the covariates of both models. Variables with the same name and different values get a presence_ or count_ prefix.

Usage

hurdle_impute(presence, count)
hurdle_impute(presence, count)

Arguments

`presence`	the `rawImputed` object for the presence.
`count`	the `rawImputed` object for counts.

Impute a dataset

Description

Impute a dataset

Usage

impute(model, ..., extra, n_imp = 19)

## S4 method for signature 'ANY'
impute(model, ..., extra, n_imp = 19)

## S4 method for signature 'glmerMod'
impute(model, data, ..., extra, n_imp)

## S4 method for signature 'maybeInla'
impute(
  model,
  ...,
  seed = 0L,
  num_threads = NULL,
  parallel_configs = TRUE,
  extra,
  n_imp = 19
)

## S4 method for signature 'lm'
impute(model, data, ..., extra, n_imp)
impute(model, ..., extra, n_imp = 19)

## S4 method for signature 'ANY'
impute(model, ..., extra, n_imp = 19)

## S4 method for signature 'glmerMod'
impute(model, data, ..., extra, n_imp)

## S4 method for signature 'maybeInla'
impute(
  model,
  ...,
  seed = 0L,
  num_threads = NULL,
  parallel_configs = TRUE,
  extra,
  n_imp = 19
)

## S4 method for signature 'lm'
impute(model, data, ..., extra, n_imp)

Arguments

`model`	model to impute the dataset
`...`	other arguments. See details
`extra`	a `data.frame` with extra observations not used in the model. They will be added in subsequent analyses.
`n_imp`	the number of imputations. Defaults to `19`.
`data`	The dataset holding both the observed and the missing values
`seed`	See the same argument in `INLA::inla.qsample()` for further information. In order to produce reproducible results, you ALSO need to make sure the RNG in R is in the same state, see the example in `INLA::inla.posterior.sample()`. When seed is non-zero, `num_threads` is forced to `"1:1"` and `parallel_configs` is set to `FALSE`, since parallel sampling would not produce a reproducible sequence of pseudo-random numbers.
`num_threads`	The number of threads to use in the format `"A:B"` defining the number threads in the outer (`A`) and inner (`B`) layer for nested parallelism. `⁠A "0"⁠` will be replaced intelligently. `seed != 0` requires serial computations.
`parallel_configs`	Logical. If TRUE and not on Windows, then try to run each configuration in parallel (not Windows) using `A` threads (see `num_threads`), where each of them is using `B:0` threads.

Examples

dataset <- generate_data(n_year = 10, n_site = 50, n_run = 1)
dataset$Count[sample(nrow(dataset), 50)] <- NA
model <- lm(Count ~ Year + factor(Period) + factor(Site), data = dataset)
impute(model, dataset)
dataset <- generate_data(n_year = 10, n_site = 50, n_run = 1)
dataset$Count[sample(nrow(dataset), 50)] <- NA
model <- lm(Count ~ Year + factor(Period) + factor(Site), data = dataset)
impute(model, dataset)

The `maybeInla` class

Description

A superclass holding either NULL or an object of the inla class.

Generate missing data at random

Description

The observed values will be either equal to the counts or missing. The probability of missing is the inverse of the counts + 1.

Usage

missing_at_random(
  dataset,
  proportion = 0.25,
  count_variable = "Count",
  observed_variable = "Observed"
)
missing_at_random(
  dataset,
  proportion = 0.25,
  count_variable = "Count",
  observed_variable = "Observed"
)

Arguments

`dataset`	A dataset to a the observation with missing data.
`proportion`	The proportion of observations that will be missing.
`count_variable`	The name of the variable holding the counts.
`observed_variable`	The name of the variable holding the observed values = either count or missing.

Generate missing data depending on the counts

Description

The observed values will be either equal to the counts or missing. The probability of missing is the inverse of the counts + 1.

Usage

missing_current_count(
  dataset,
  proportion = 0.25,
  count_variable = "Count",
  observed_variable = "Observed"
)
missing_current_count(
  dataset,
  proportion = 0.25,
  count_variable = "Count",
  observed_variable = "Observed"
)

Arguments

`dataset`	A dataset to a the observation with missing data.
`proportion`	The proportion of observations that will be missing.
`count_variable`	The name of the variable holding the counts.
`observed_variable`	The name of the variable holding the observed values = either count or missing.

Generate missing data based on the observed patterns in the real dataset.

Description

The observed values will be either equal to the counts or missing. The probability of missing is the inverse of the counts + 1.

Usage

missing_observed(
  dataset,
  count_variable = "Count",
  observed_variable = "Observed",
  site_variable = "Site",
  year_variable = "Year",
  period_variable = "Period"
)
missing_observed(
  dataset,
  count_variable = "Count",
  observed_variable = "Observed",
  site_variable = "Site",
  year_variable = "Year",
  period_variable = "Period"
)

Arguments

`dataset`	A dataset to a the observation with missing data.
`count_variable`	The name of the variable holding the counts.
`observed_variable`	The name of the variable holding the observed values = either count or missing.
`site_variable`	The name of the variable holding the sites.
`year_variable`	The name of the variable holding the years.
`period_variable`	The name of the variable holding the period.

Generate missing data mimicking choices made by volunteers.

Description

The observed values will be either equal to the counts or missing. The probability of missing is the inverse of the counts + 1.

Usage

missing_volunteer(
  dataset,
  proportion = 0.25,
  count_variable = "Count",
  observed_variable = "Observed",
  year_variable = "Year",
  site_variable = "Site",
  max_count = 100
)
missing_volunteer(
  dataset,
  proportion = 0.25,
  count_variable = "Count",
  observed_variable = "Observed",
  year_variable = "Year",
  site_variable = "Site",
  max_count = 100
)

Arguments

`dataset`	A dataset to a the observation with missing data.
`proportion`	The proportion of observations that will be missing.
`count_variable`	The name of the variable holding the counts.
`observed_variable`	The name of the variable holding the observed values = either count or missing.
`year_variable`	The name of the variable holding the years.
`site_variable`	The name of the variable holding the sites.
`max_count`	The maximum count.

Model an imputed dataset

Description

Model an imputed dataset

Usage

model_impute(
  object,
  model_fun,
  rhs,
  model_args = list(),
  extractor,
  extractor_args = list(),
  filter = list(),
  mutate = list(),
  ...,
  timeout = 600
)

## S4 method for signature 'ANY'
model_impute(
  object,
  model_fun,
  rhs,
  model_args = list(),
  extractor,
  extractor_args = list(),
  filter = list(),
  mutate = list(),
  ...,
  timeout = 600
)

## S4 method for signature 'aggregatedImputed'
model_impute(
  object,
  model_fun,
  rhs,
  model_args = list(),
  extractor,
  extractor_args = list(),
  filter = list(),
  mutate = list(),
  ...,
  timeout = 600
)
model_impute(
  object,
  model_fun,
  rhs,
  model_args = list(),
  extractor,
  extractor_args = list(),
  filter = list(),
  mutate = list(),
  ...,
  timeout = 600
)

## S4 method for signature 'ANY'
model_impute(
  object,
  model_fun,
  rhs,
  model_args = list(),
  extractor,
  extractor_args = list(),
  filter = list(),
  mutate = list(),
  ...,
  timeout = 600
)

## S4 method for signature 'aggregatedImputed'
model_impute(
  object,
  model_fun,
  rhs,
  model_args = list(),
  extractor,
  extractor_args = list(),
  filter = list(),
  mutate = list(),
  ...,
  timeout = 600
)

Arguments

`object`	The imputed dataset.
`model_fun`	The function to apply on each imputation set. Or a string with the name of the function. Include the package name when the function is not in one of the base R packages. For example: `"glm"` or `"INLA::inla"`.
`rhs`	The right hand side of the model.
`model_args`	An optional list of arguments to pass to the model function.
`extractor`	A function which return a `matrix` or `data.frame`. The first column should contain the estimate, the second the standard error of the estimate.
`extractor_args`	An optional list of arguments to pass to the `extractor` function.
`filter`	An optional argument to filter the aggregated dataset. Either a function which takes the `Covariate` slot as an argument. Or a list which will be passed to the `.dots` argument of `dplyr::filter()`. You can filter on the covariates in the aggregated dataset. Besides those you can also filter on `Imputation_min` and `Imputation_max`. These variables represent the lowest and highest value of the imputations per row in the data.
`mutate`	An optional argument to alter the aggregated dataset. Will be passed to the `.dots` argument of `dplyr::mutate()`. This is mainly useful for simple conversions, e.g. factors to numbers and vice versa.
`...`	currently ignored.
`timeout`	Maximum duration allowed for fitting a single imputation model in seconds. Defaults to `600` seconds (10 minutes).

Examples

dataset <- generate_data(n_year = 10, n_site = 50, n_run = 1)
dataset$Count[sample(nrow(dataset), 50)] <- NA
model <- lm(Count ~ Year + factor(Period) + factor(Site), data = dataset)
imputed <- impute(data = dataset, model = model)
aggr <- aggregate_impute(imputed, grouping = c("Year", "Period"), fun = sum)
extractor <- function(model) {
  summary(model)$coefficients[, c("Estimate", "Std. Error")]
}
model_impute(
  object = aggr,
  model_fun = lm,
  rhs = "0 + factor(Year)",
  extractor = extractor
)
dataset <- generate_data(n_year = 10, n_site = 50, n_run = 1)
dataset$Count[sample(nrow(dataset), 50)] <- NA
model <- lm(Count ~ Year + factor(Period) + factor(Site), data = dataset)
imputed <- impute(data = dataset, model = model)
aggr <- aggregate_impute(imputed, grouping = c("Year", "Period"), fun = sum)
extractor <- function(model) {
  summary(model)$coefficients[, c("Estimate", "Std. Error")]
}
model_impute(
  object = aggr,
  model_fun = lm,
  rhs = "0 + factor(Year)",
  extractor = extractor
)

The `rawImputed` class Holds a dataset and imputed values

Description

The rawImputed class Holds a dataset and imputed values

Slots

Data: A data.frame with the data.
Response: A character holding the name of the response variable.
Minimum: An optional character holding the name of the variable with the minimum.
Imputation: A matrix with imputed values.
Extra: A data.frame with extra data to add to the imputations. This data is not used in the imputation model. It must contain the same variables as the original data.

The observation pattern in the Flemish waterfowl dataset

Description

Data for fig 1 and 2 in Onkelinx et al

Usage

data(waterfowl)
data(waterfowl)

Format

A data frame with 77157 rows and 5 variables

Details

Site Site ID.
Winter Winter ID.
Period ID of the month.
Species Number of observed species.
Birds Total number of birds.

Package 'multimput'

Help Index

Aggregate an imputed dataset

Description

Usage

Arguments

Examples

The aggregatedImputed class Holds an aggregated imputation data set

Description

Slots

Generate simulated data

Description

Usage

Arguments

Value

Combine two models into a hurdle model

Description

Usage

Arguments

Impute a dataset

Description

Usage

Arguments

Examples

The maybeInla class

Description

Generate missing data at random

Description

Usage

Arguments

Generate missing data depending on the counts

Description

Usage

Arguments

Generate missing data based on the observed patterns in the real dataset.

Description

Usage

Arguments

Generate missing data mimicking choices made by volunteers.

Description

Usage

Arguments

Model an imputed dataset

Description

Usage

Arguments

Examples

The rawImputed class Holds a dataset and imputed values

Description

Slots

The observation pattern in the Flemish waterfowl dataset

Description

Usage

Format

Details

The `aggregatedImputed` class Holds an aggregated imputation data set

The `maybeInla` class

The `rawImputed` class Holds a dataset and imputed values