Constructor for bivariate range object

Description

Constructor for bivariate range object

Usage

bivariate_range()

Format

An object of class RangeBivariate (inherits from Range, ggproto, gg) of length 2.

Constructor for bivariate scale object

Description

Constructor for bivariate scale object

Usage

bivariate_scale(
  aesthetics,
  palette,
  name = waiver(),
  breaks = waiver(),
  labels = waiver(),
  limits = NULL,
  rescaler = scales::rescale,
  oob = scales::censor,
  expand = waiver(),
  na.value = NA_real_,
  trans = "identity",
  guide = "none",
  super = ScaleBivariate,
  scale_name = "bivariate_scale"
)

Arguments

Format

An object of class ScaleBivariate (inherits from Scale, ggproto, gg) of length 15.

acceptRate

Description

Plots the acceptance rate of trees from a BART model.

Usage

acceptRate(trees)

Arguments

Value

A ggplot object plot of acceptance rate.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
# Load the dbarts package to access the bart function
library(dbarts)
# Get Data
df <- na.omit(airquality)
# Create Simple dbarts Model For Regression:
set.seed(1701)
dbartModel <- bart(df[2:6], df[,1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)
# Tree Data
trees_data <- extractTreeData(model = dbartModel, data = df)
acceptRate(trees = trees_data)}

bartClassifDiag

Description

Displays a selection of diagnostic plots for a BART model.

Usage

bartClassifDiag(
  model,
  data,
  response,
  threshold = "Youden",
  pNorm = FALSE,
  showInterval = TRUE,
  combineFactors = FALSE
)

Arguments

Value

A selection of diagnostic plots

bartDiag

Description

Displays a selection of diagnostic plots for a BART model.

Usage

bartDiag(
  model,
  data,
  response,
  burnIn = 0,
  threshold = "Youden",
  pNorm = FALSE,
  showInterval = TRUE,
  combineFactors = FALSE
)

Arguments

Value

A selection of diagnostic plots.

Examples

# For Regression
# Generate Friedman data
fData <- function(n = 200, sigma = 1.0, seed = 1701, nvar = 5) {
  set.seed(seed)
  x <- matrix(runif(n * nvar), n, nvar)
  colnames(x) <- paste0("x", 1:nvar)
  Ey <- 10 * sin(pi * x[, 1] * x[, 2]) + 20 * (x[, 3] - 0.5)^2 + 10 * x[, 4] + 5 * x[, 5]
  y <- rnorm(n, Ey, sigma)
  data <- as.data.frame(cbind(x, y))
  return(data)
}
f_data <- fData(nvar = 10)
x <- f_data[, 1:10]
y <- f_data$y

# Create dbarts model
library(dbarts)
set.seed(1701)
dbartModel <- bart(x, y, ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

bartDiag(model = dbartModel, response = "y", burnIn = 100, data = f_data)


# For Classification
data(iris)
iris2 <- iris[51:150, ]
iris2$Species <- factor(iris2$Species)

# Create dbarts model
dbartModel <- bart(iris2[, 1:4], iris2[, 5], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

bartDiag(model = dbartModel, data = iris2, response = iris2$Species)

bartRegrDiag

Description

Displays a selection of diagnostic plots for a BART model.

Usage

bartRegrDiag(model, response, burnIn = 0, data, combineFactors = FALSE)

Arguments

Value

A selection of diagnostic plots

Cluster Trees by Variable

Description

Reorders a list of tree structures based on the clustering of variables within each tree.

Usage

clusterTrees(tree_list)

Arguments

Value

A list of trees reordered based on the clustering of variables.

Update Dummy Variable Names

Description

This function updates the 'var' column in the 'structure' component of the 'trees' list, replacing dummy variable names derived from factor variables with their original factor variable names.

Usage

combineDummy(trees)

Arguments

Details

The function first identifies factor variables in 'trees$data', then checks each entry in 'trees$structure$var' for matches with these factor variables. If a match is found, indicating a dummy variable, the entry is replaced with the original factor variable name.

Value

The modified 'trees' list with updated 'var' column entries in 'trees$structure'.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
# Load the dbarts package to access the bart function
 library(dbarts)
 # Create Simple dbarts Model with Dummies
 set.seed(1701)
 dbartModel <- bart(iris[2:5], iris[,1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = iris)
 combined_trees <- combineDummy(trees = trees_data)
}

extractTreeData

Description

Creates a list of all tree attributes for a model created by either the BART, dbarts or bartMachine packages.

Usage

extractTreeData(model, data)

Arguments

Value

A list containing the extracted and processed tree data. This list includes:

1. Tree Data Frame: A data frame containing tree attributes.

2. Variable Name: The names of the variables used in building the model.

3. nMCMC: The total number of iterations (posterior draws) after burn-in.

4. nTree: The total number of trees grown in the sum-of-trees model.

5. nVar: The total number of covariates used in the model.

The object created by the 'extractTreeData' function encompasses these elements, facilitating detailed analysis and visualisation of BART model components.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
# Load the dbarts package to access the bart function
library(dbarts)
# Get Data
df <- na.omit(airquality)
# Create Simple dbarts Model For Regression:
set.seed(1701)
dbartModel <- bart(df[2:6], df[,1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)
# Tree Data
trees_data <- extractTreeData(model = dbartModel, data = df)
}

Generate Child and Parent Node Relationships

Description

Populates 'childLeft', 'childRight', and 'parent' columns in the dataset to establish parent-child relationships between nodes based on tree structure.

Usage

getChildren(data)

Arguments

Value

The modified data frame with 'childLeft', 'childRight', and 'parent' columns added, detailing the tree's parent-child node relationships.

Examples

data("tree_data_example")
# Create Terminal Column
tree_data_example <- transform(tree_data_example,
                               terminal = ifelse(is.na(var),
                               TRUE,
                               FALSE))
# Get depths
depthList <- lapply(split(tree_data_example, ~treeNum + iteration),
                    function(x) cbind(x, depth = node_depth(x)-1))
# Turn into data frame
tree_data_example <- dplyr::bind_rows(depthList, .id = "list_id")
# Add node number sequntially
tree_data_example$node <- with(tree_data_example,
                               ave(seq_along(iteration),
                               list(iteration, treeNum),
                               FUN = seq_along))
# get children
getChildren(data = tree_data_example)

Get Observations Falling into Each Node

Description

This function determines which observations from a given dataset fall into which nodes of a tree, based on a tree structure defined in 'treeData'. The treeData object must include 'iteration', 'treeNum', 'var', and 'splitValue' columns.

Usage

getObservations(data, treeData)

Arguments

Value

A modified version of 'treeData' that includes two new columns: 'obsNode' and 'noObs'. 'obsNode' lists the observations falling into each node, and 'noObs' provides the count of observations for each node.

Examples

data("tree_data_example")
# Create Terminal Column
tree_data_example <- transform(tree_data_example,
                              terminal = ifelse(is.na(var),
                              TRUE,
                               FALSE))
# Create Split Value Column
tree_data_example <- transform(tree_data_example,
                               splitValue = ifelse(terminal == FALSE,
                               value,
                               NA_integer_))
# get the observations
getObservations(data = input_data, treeData = tree_data_example)

Determines the stump color for a legend based on its mean value

Description

This function is internal and is used to compute the color of a stump for the purpose of legend display, based on the mean value relative to specified limits.

Usage

get_stump_colour_for_legend(lims, mean_value, palette)

Arguments

Value

A character string specifying the color corresponding to the mean value.

Colourfan guide

Description

Colourfan guide

Usage

guide_colourfan(
  title = waiver(),
  title.x.position = "top",
  title.y.position = "right",
  title.theme = NULL,
  title.hjust = 0.5,
  title.vjust = NULL,
  label = TRUE,
  label.theme = NULL,
  barwidth = NULL,
  barheight = NULL,
  nbin = 32,
  reverse = FALSE,
  order = 0,
  available_aes = c("colour", "color", "fill"),
  ...
)

guide_colorfan(
  title = waiver(),
  title.x.position = "top",
  title.y.position = "right",
  title.theme = NULL,
  title.hjust = 0.5,
  title.vjust = NULL,
  label = TRUE,
  label.theme = NULL,
  barwidth = NULL,
  barheight = NULL,
  nbin = 32,
  reverse = FALSE,
  order = 0,
  available_aes = c("colour", "color", "fill"),
  ...
)

Arguments

Value

A 'grob' object representing a color fan. This 'grob' can be added to a grid-based plot or a ggplot2 object to visualize a range of colors in a fan-like structure. Each segment of the fan corresponds to a color specified in the 'colours' parameter, allowing for an intuitive representation of color gradients or palettes.

input_data

Description

Small example of Friedman data following the formula:

y = 10 \sin(\pi x_1 x_2) + 20 (x_3 - 0.5)^2 + 10 x_4 + 5 x_5 + e

Usage

input_data

Format

A data frame with 10 rows and 6 columns:

x1

Covariate

x2

Covariate

x3

Covariate

x4

Covariate

x5

Covariate

y

Response

...

localProcedure

Description

A variable selection approach performed by permuting the response.

Usage

localProcedure(
  model,
  data,
  response,
  numRep = 10,
  numTreesRep = NULL,
  alpha = 0.5,
  shift = FALSE
)

Arguments

Value

A variable selection plot using the local procedure method.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
# Load the dbarts package to access the bart function
library(dbarts)

# Get Data
df <- na.omit(airquality)
# Create Simple dbarts Model For Regression:
set.seed(1701)
dbartModel <- bart(df[2:6], df[,1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)
localProcedure(model = dbartModel,
               data = df,
               numRep = 5,
               numTreesRep = 5,
               alpha = 0.5,
               shift = FALSE)
}

mdsBart

Description

Multi-dimensional Scaling Plot of proximity matrix from a BART model.

Usage

mdsBart(
  trees,
  data,
  target,
  response,
  plotType = "rows",
  showGroup = TRUE,
  level = 0.95
)

Arguments

Value

For this function, the MDS coordinates are calculated for each iteration. Procrustes method is then applied to align each of the coordinates to a target set of coordinates. The returning result is then a clustered average of each point.

Examples

if (requireNamespace("dbarts", quietly = TRUE)) {
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6],
   df[, 1],
   ntree = 5,
   keeptrees = TRUE,
   nskip = 10,
   ndpost = 10
 )
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 # Cretae Porximity Matrix
 bmProx <- proximityMatrix(
   trees = trees_data,
   reorder = TRUE,
   normalize = TRUE,
   iter = 1
 )
 # MDS plot
 mdsBart(
   trees = trees_data, data = df, target = bmProx,
   plotType = "interactive", level = 0.25, response = "Ozone"
 )
}

Calculate Node Depths in a Tree Data Frame

Description

Computes the depth of each node in a given tree data frame, assuming a binary tree structure. Requires the tree data frame to contain a logical column 'terminal' indicating terminal nodes.

Usage

node_depth(tree)

Arguments

Value

A vector of depths corresponding to each node in the tree.

Examples

data("tree_data_example")
# Create Terminal Column
tree_data_example <- transform(tree_data_example, terminal = ifelse(is.na(var), TRUE, FALSE))
# Get depths
depthList <- lapply(split(tree_data_example, ~treeNum + iteration),
                    function(x) cbind(x, depth = node_depth(x)-1))
# Turn into data frame
tree_data_example <- dplyr::bind_rows(depthList, .id = "list_id")

Variance suppressing uncertainty palette

Description

Returns a palette function that turns 'v' (value) and 'u' (uncertainty) (both between 0 and 1) into colors.

Usage

pal_vsup(
  values,
  unc_levels = 4,
  max_light = 0.9,
  max_desat = 0,
  pow_light = 0.8,
  pow_desat = 1
)

Arguments

Value

A function that takes two parameters, 'v' (value) and 'u' (uncertainty), both expected to be in the range of 0 to 1, and returns a color. This color is determined by the specified 'values' colors at minimum uncertainty, and modified according to the given 'v' and 'u' parameters to represent uncertainty by adjusting lightness and saturation. The resulting function is useful for creating color palettes that can encode both value and uncertainty in visualizations.

permVimp

Description

A variable selection approach which creates a null model by permuting the response, rebuilding the model, and calculating the inclusion proportion (IP) on the null model. The final result displayed is the original model's IP minus the null IP.

Usage

permVimp(model, data, response, numTreesPerm = NULL, plotType = "barplot")

Arguments

Value

A variable selection plot.

Examples

if (requireNamespace("dbarts", quietly = TRUE)) {
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6],
   df[, 1],
   ntree = 5,
   keeptrees = TRUE,
   nskip = 10,
   ndpost = 10
 )
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 permVimp(model = dbartModel, data = df, response = 'Ozone', numTreesPerm = 2, plotType = 'point')
}

permVint

Description

A variable interaction evaluation which creates a null model by permuting the response, rebuilding the model, and calculating the inclusion proportion (IP) of adjacent splits on the null model. The final result displayed is the original model's IP minus the null IP.

Usage

permVint(model, data, trees, response, numTreesPerm = NULL, top = NULL)

Arguments

Value

A variable interaction plot. Note that for a dbarts fit, due to the internal workings of dbarts, the null model is hard-coded to 20 trees, a burn-in of 100, and 1000 iterations. Both a BART and bartMachine null model will extract the identical parameters from the original model.

plotProximity

Description

Plot a proximity matrix

Usage

plotProximity(
  matrix,
  pal = rev(colorspace::sequential_hcl(palette = "Blues 2", n = 100)),
  limit = NULL
)

Arguments

Value

A plot of proximity values.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6],
   df[, 1],
   ntree = 5,
   keeptrees = TRUE,
   nskip = 10,
   ndpost = 10)
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 # Create Proximity Matrix
 mProx <- proximityMatrix(trees = trees_data, reorder = TRUE, normalize = TRUE, iter = 1)
 # Plot
 plotProximity(matrix = mProx)
}

plotSingleTree

Description

Plots individual trees.

Usage

plotSingleTree(trees, iter = 1, treeNo = 1, plotType = "icicle")

Arguments

Value

A plot of an individual tree

Examples

if (requireNamespace("dbarts", quietly = TRUE)) {
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6],
   df[, 1],
   ntree = 5,
   keeptrees = TRUE,
   nskip = 10,
   ndpost = 10
 )
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 plotSingleTree(trees = trees_data, iter = 1, treeNo = 1)
}

Plot Trees with Customisations

Description

This function plots trees from a list of tidygraph objects. It allows for various customisations such as fill colour based on node response or value, node size adjustments, and color palettes.

Usage

plotTrees(
  trees,
  iter = NULL,
  treeNo = NULL,
  fillBy = NULL,
  sizeNodes = FALSE,
  removeStump = FALSE,
  selectedVars = NULL,
  pal = rev(colorRampPalette(c("steelblue", "#f7fcfd", "orange"))(5)),
  center_Mu = TRUE,
  cluster = NULL
)

Arguments

Value

A ggplot object representing the plotted trees with the specified customisations.

Examples

if (requireNamespace("dbarts", quietly = TRUE)) {
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6],
   df[, 1],
   ntree = 5,
   keeptrees = TRUE,
   nskip = 10,
   ndpost = 10
 )
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 plotTrees(trees = trees_data, fillBy = 'response', sizeNodes = TRUE)
}

print.hideHelper

Description

This function hides parts from the print out but are still accessible via indexing.

Usage

## S3 method for class 'hideHelper1'
print(x, ...)

Arguments

Value

No return value; this function is called for its side effect of printing a formatted summary of the tree data frame. It displays parts of the data frame, such as the tree structure and various counts (like number of MCMC iterations, number of trees, and number of variables), while keeping the complete data accessible via indexing.

proximityMatrix

Description

Creates a matrix of proximity values.

Usage

proximityMatrix(trees, nRows, normalize = TRUE, reorder = TRUE, iter = NULL)

Arguments

Value

A matrix containing proximity values.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 # Create Proximity Matrix
 mProx <- proximityMatrix(trees = trees_data, reorder = TRUE, normalize = TRUE, iter = 1)
}

Sort Trees by Maximum Depth

Description

Sort Trees by Maximum Depth

Usage

sort_trees_by_depthMax(tree_list)

Arguments

Value

Sorted list of 'tbl_graph' trees by decreasing maximum depth.

splitDensity

Description

Density plots of the split value for each variable.

Usage

splitDensity(
  trees,
  data,
  bandWidth = NULL,
  panelScale = NULL,
  scaleFactor = NULL,
  display = "histogram"
)

Arguments

Value

A faceted group of density plots

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 splitDensity(trees = trees_data, data = df, display = 'ridge')
}

Generate Terminal Node Indicator

Description

Adds a boolean 'terminal' column to the dataset indicating whether each node is terminal.

Usage

terminalFunction(data)

Arguments

Value

The modified data frame with an additional 'terminal' column.

Train range for bivariate scale

Description

Train range for bivariate scale

Usage

train_bivariate(new, existing = NULL)

Arguments

Value

A tibble containing two columns, 'range1' and 'range2', each representing the trained continuous range based on the new and existing data. This function is used to update or define the scales of a bivariate analysis by considering both new input data and any existing range specifications.

Plot Frequency of Tree Structures

Description

Generates a bar plot showing the frequency of different tree structures represented in a list of tree graphs. Optionally, it can filter to show only the top N trees and handle stump trees specially.

Usage

treeBarPlot(trees, iter = NULL, topTrees = NULL, removeStump = FALSE)

Arguments

Details

This function processes a list of tree structures to compute the frequency of each unique structure, represented by a bar plot. It has options to exclude stump trees (trees with no edges) and to limit the plot to the top N most frequent structures.

Value

A 'ggplot' object representing the bar plot of tree frequencies.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 plot <- treeBarPlot(trees = trees_data, topTrees = 3, removeStump = TRUE)
}

treeDepth

Description

A plot of tree depth over iterations.

Usage

treeDepth(trees)

Arguments

Value

A plot of average tree depths over iteration

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 treeDepth(trees = trees_data)
}

Generate a List of Tree Structures from BART Model Output

Description

This function takes a dataframe of trees, which is output from a BART model, and organizes it into a list of tree structures. It allows for filtering based on iteration number, tree number, and optionally reordering based on the maximum depth of nodes or variables.

Usage

treeList(trees, iter = NULL, treeNo = NULL)

Arguments

Value

A list of tidygraph objects, each representing the structure of a tree. Each tidygraph object includes node and edge information necessary for visualisation.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 library(ggplot2)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 trees_list <- treeList(trees_data)
}

treeNodes

Description

A plot of number of nodes over iterations.

Usage

treeNodes(trees)

Arguments

Value

A plot of tree number of nodes over iterations.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 treeNodes(trees = trees_data)
}

tree_data_example

Description

Small example of tree data, like that obtained when using 'extractTreeData()' function.

Usage

tree_data_example

Format

A data frame with 14 rows and 4 columns representing the structure of trees:

var

Variable name used for splitting.

value

The value in a node (i.e., either the split value or leaf value).

iteration

Iteration Number.

treeNum

Tree Number in the iteration.

...

Transform tree data into a structured dataframe

Description

This function takes raw data and a tree structure, then processes it to form a detailed and structured dataframe. The data is transformed to indicate terminal nodes, calculate leaf values, and determine split values. It then assigns labels, calculates node depth, and establishes hierarchical relationships within the tree. Additional metadata about the tree, such as maximum depth, parent and child node relationships, and observation nodes are also included. The final dataframe is organized and enriched with necessary attributes for further analysis.

Usage

tree_dataframe(data, trees, response = NULL)

Arguments

Value

A list containing a detailed dataframe of the tree structure ('structure') with added information such as node depth, parent and child nodes, and observational data, along with meta-information about the tree like variable names ('varNames'), number of MCMC iterations ('nMCMC'), number of trees ('nTree'), and number of variables ('nVar').

Examples

data("input_data")
data("tree_data_example")
my_trees <- tree_dataframe(data = input_data, trees = tree_data_example, response = "y")

vimpBart

Description

A matrix with nMCMC rows with each variable as a column. Each row represents an MCMC iteration. For each variable, the total count of the number of times that variable is used in a tree is given.

Usage

vimpBart(trees, type = "prop")

Arguments

Value

A matrix of importance values

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 vimpBart(trees_data, type = 'prop')
 }

vimpPlot

Description

Plot the variable importance for a BART model with the 25 quantile.

Usage

vimpPlot(trees, type = "prop", plotType = "barplot", metric = "median")

Arguments

Value

A plot of variable importance.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 vimpPlot(trees = trees_data, plotType = 'point')
}

vintPlot

Description

Plot the pair-wise variable interactions inclusion porportions for a BART model with the 25

Usage

vintPlot(trees, plotType = "barplot", top = NULL)

Arguments

Value

A plot of variable importance.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 vintPlot(trees = trees_data, top = 5)
}

viviBart

Description

Returns a list containing a dataframe of variable importance summaries and a dataframe of variable interaction summaries.

Usage

viviBart(trees, out = "vivi")

Arguments

Value

A list of dataframes of VIVI summaries.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 viviBart(trees = trees_data, out = 'vivi')
 }

viviBartMatrix

Description

Returns a matrix or list of matrices. If type = 'standard' a matrix filled with vivi values is returned. If type = 'vsup' two matrices are returned. One with the actual values and another matrix of uncertainty values. If type = 'quantiles', three matrices are returned. One for the 25

Usage

viviBartMatrix(
  trees,
  type = "standard",
  metric = "propMean",
  metricError = "CV",
  reorder = FALSE
)

Arguments

Value

A heatmap plot showing variable importance on the diagonal and variable interaction on the off-diagonal.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)

 # VSUP Matrix
 vsupMat <- viviBartMatrix(trees = trees_data,
                           type = 'vsup',
                           metric = 'propMean',
                            metricError = 'CV')
 }

viviBartPlot

Description

Plots a Heatmap showing variable importance on the diagonal and variable interaction on the off-diagonal with uncertainty included.

Usage

viviBartPlot(
  matrix,
  intPal = NULL,
  impPal = NULL,
  intLims = NULL,
  impLims = NULL,
  uncIntLims = NULL,
  uncImpLims = NULL,
  unc_levels = 4,
  max_desat = 0.6,
  pow_desat = 0.2,
  max_light = 0.6,
  pow_light = 1,
  angle = 0,
  border = FALSE,
  label = NULL
)

Arguments

Value

Either a heatmap, VSUP, or quantile heatmap plot.

Examples

if(requireNamespace("dbarts", quietly = TRUE)){
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6], df[, 1], ntree = 5, keeptrees = TRUE, nskip = 10, ndpost = 10)

 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)

 # VSUP Matrix
 vsupMat <- viviBartMatrix(trees = trees_data,
                           type = 'vsup',
                           metric = 'propMean',
                           metricError = 'CV')
 # Plot
 viviBartPlot(vsupMat, label = 'CV')
 }