compare_plot {povmap}

Shows Plots for the Comparison of Estimates

Description

Function compare_plot is a generic function used to produce plots comparing point and existing MSE/CV estimates of direct and model-based estimation for all indicators or a selection of indicators.

Methods compare_plot.direct, compare_plot.ebp and compare_plot.fh produce plots comparing point and existing MSE/CV estimates of direct and model-based estimation for all indicators or a selection of indicators for objects of type "emdi". The direct and model-based point estimates are compared by a scatter plot and a line plot for each selected indicator. If the input arguments MSE and CV are set to TRUE, two extra plots are created, respectively: the MSE/CV estimates of the direct and model-based estimates are compared by boxplots and scatter plots.

Usage

compare_plot(
  model,
  direct,
  indicator = "all",
  MSE = FALSE,
  CV = FALSE,
  label = "orig",
  color = c("blue", "lightblue3"),
  shape = c(16, 16),
  line_type = c("solid", "solid"),
  gg_theme = NULL,
  ...
)

## S3 method for class 'direct'
compare_plot(
  model = NULL,
  direct = NULL,
  indicator = "all",
  MSE = FALSE,
  CV = FALSE,
  label = "orig",
  color = c("blue", "lightblue3"),
  shape = c(16, 16),
  line_type = c("solid", "solid"),
  gg_theme = NULL,
  ...
)

## S3 method for class 'ebp'
compare_plot(
  model = NULL,
  direct = NULL,
  indicator = "all",
  MSE = FALSE,
  CV = FALSE,
  label = "orig",
  color = c("blue", "lightblue3"),
  shape = c(16, 16),
  line_type = c("solid", "solid"),
  gg_theme = NULL,
  ...
)

## S3 method for class 'fh'
compare_plot(
  model = NULL,
  direct = NULL,
  indicator = "all",
  MSE = FALSE,
  CV = FALSE,
  label = "orig",
  color = c("blue", "lightblue3"),
  shape = c(16, 16),
  line_type = c("solid", "solid"),
  gg_theme = NULL,
  ...
)

Arguments

model

a model object of type "emdi", either "ebp" or "fh", representing point and MSE estimates.

direct

an object of type "direct","emdi", representing point and MSE estimates. If the input argument model is of type "ebp", direct is required. If the input argument model is of type "fh", the direct component is already included in the input argument model.

indicator

optional character vector that selects which indicators shall be returned: (i) all calculated indicators ("all"); (ii) each indicator name: "Mean", "Quantile_10", "Quantile_25", "Median", "Quantile_75", "Quantile_90", "Head_Count", "Poverty_Gap", "Gini", "Quintile_Share" or the function name/s of "custom_indicator/s"; (iii) groups of indicators: "Quantiles", "Poverty", "Inequality" or "Custom". If two of these groups are selected, only the first one is returned. Note, additional custom indicators can be defined as argument for the EBP approaches (see also ebp) and do not appear in groups of indicators even though these might belong to one of the groups. If the model argument is of type "fh", indicator can be set to "all", "Direct", FH", or "FH_Bench" (if emdi object is overwritten by function benchmark). Defaults to "all".

MSE

optional logical. If TRUE, the MSE estimates of the direct and model-based estimates are compared via boxplots and scatter plots.

CV

optional logical. If TRUE, the coefficient of variation estimates of the direct and model-based estimates are compared via boxplots and scatter plots.

label

argument that enables to customize title and axis labels. There are three options to label the evaluation plots: (i) original labels ("orig"), (ii) axis labels but no title ("no_title"), (iii) neither axis labels nor title ("blank").

color

a vector with two elements. The first color determines the color for the regression line in the scatter plot and the color for the direct estimates in the remaining plots. The second color specifies the color of the intersection line in the scatter plot and the color for the model-based estimates in the remaining plots. Defaults to c("blue", "lightblue3").

shape

a numeric vector with two elements. The first shape determines the shape of the points in the scatterplot and the shape of the points for the direct estimates in the remaining plots. The second shape determines the shape for the points for the model-based estimates. The options are numbered from 0 to 25. Defaults to c(16, 16).

line_type

a character vector with two elements. The first line type determines the line type for the regression line in the scatter plot and the line type for the direct estimates in the remaining plots. The second line type specifies the line type of the intersection line in the scatter plot and the line type for the model-based estimates in the remaining plots. The options are: "twodash", "solid", "longdash", "dotted", "dotdash", "dashed" and "blank". Defaults to c("solid", "solid").

gg_theme

theme list from package ggplot2. For using this argument, package ggplot2 must be loaded via library(ggplot2). See also Example 2.

...

further arguments passed to or from other methods.

Details

Since all of the comparisons need a direct estimator, the plots are only created for in-sample domains.

Value

Plots comparing direct and model-based estimators for each selected indicator obtained by ggplot.

A scatter plot and a line plot comparing direct and model-based estimators for each selected indicator obtained by ggplot. If the input arguments MSE and CV are set to TRUE two extra plots are created, respectively: the MSE/CV estimates of the direct and model-based estimates are compared by boxplots and scatter plots.

See Also

emdiObject, direct, ebp, fh

Examples

# Examples for comparisons of direct estimates and models of type ebp

# Loading data - population and sample data
data("eusilcA_pop")
data("eusilcA_smp")

# Generation of two emdi objects
emdi_model <- ebp(
  fixed = eqIncome ~ gender + eqsize + cash +
    self_empl + unempl_ben + age_ben + surv_ben + sick_ben + dis_ben + rent +
    fam_allow + house_allow + cap_inv + tax_adj, pop_data = eusilcA_pop,
  pop_domains = "district", smp_data = eusilcA_smp, smp_domains = "district",
  threshold = function(y) {
    0.6 * median(y)
  }, L = 50, MSE = TRUE,
  na.rm = TRUE, cpus = 1
)

emdi_direct <- direct(
  y = "eqIncome", smp_data = eusilcA_smp,
  smp_domains = "district", weights = "weight", threshold = 11161.44,
  var = TRUE, boot_type = "naive", B = 50, seed = 123, na.rm = TRUE
)

# Example 1: Receive first overview
compare_plot(model = emdi_model, direct = emdi_direct)

# Example 2: Change plot theme
library(ggplot2)
compare_plot(emdi_model, emdi_direct,
  indicator = "Median",
  gg_theme = theme(
    axis.line = element_line(size = 3, colour = "grey80"),
    plot.background = element_rect(fill = "lightblue3"),
    legend.position = "none"
  )
)

# Example for comparison of direct estimates and models of type fh

# Loading data - population and sample data
data("eusilcA_popAgg")
data("eusilcA_smpAgg")

# Combine sample and population data
combined_data <- combine_data(
  pop_data = eusilcA_popAgg,
  pop_domains = "Domain",
  smp_data = eusilcA_smpAgg,
  smp_domains = "Domain"
)

# Generation of the emdi object
fh_std <- fh(
  fixed = Mean ~ cash + self_empl, vardir = "Var_Mean",
  combined_data = combined_data, domains = "Domain",
  method = "ml", MSE = TRUE
)
# Example 3: Receive first overview
compare_plot(fh_std)

# Example 4: Compare also MSE and CV estimates
compare_plot(fh_std, MSE = TRUE, CV = TRUE)