Package 'vivid' reference manual

Title:	Variable Importance and Variable Interaction Displays
Description:	A suite of plots for displaying variable importance and two-way variable interaction jointly. Can also display partial dependence plots laid out in a pairs plot or 'zenplots' style.
Authors:	Alan Inglis [aut, cre], Andrew Parnell [aut], Catherine Hurley [aut]
Maintainer:	Alan Inglis <[email protected]>
License:	GPL (>=2)
Version:	0.2.9
Built:	2025-02-26 06:19:12 UTC
Source:	https://github.com/alaninglis/vivid

as.data.frame.vivid

Description

Takes a matrix of class vivid and turn it into a data frame containing variable names, Vimp and Vint values, and the row and column index from the original matrix.

Usage

## S3 method for class 'vivid'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)
## S3 method for class 'vivid'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

Arguments

`x`	A matrix of class 'vivid' to be converted to a data frame.
`row.names`	NULL or a character vector giving the row names for the data frame. Missing values are not allowed.
`optional`	Logical. If TRUE, setting row names and converting column names (to syntactic names: see make.names) is optional. Note that all of R's base package as.data.frame() methods use optional only for column names treatment, basically with the meaning of data.frame(*, check.names = !optional). See also the make.names argument of the matrix method.
`...`	Additional arguments to be passed to or from methods.

Value

A data frame of Vimp and Vint values and their index from the vivid matrix.

Examples

 
library(ranger)
aq <- na.omit(airquality)
aq <- aq[1:20,]# for speed
rF <- ranger(Ozone ~ ., data = aq, importance = "permutation")
myMat <- vivi(fit = rF, data = aq, response = "Ozone")
myDf <- as.data.frame(myMat)
myDf

library(ranger)
aq <- na.omit(airquality)
aq <- aq[1:20,]# for speed
rF <- ranger(Ozone ~ ., data = aq, importance = "permutation")
myMat <- vivi(fit = rF, data = aq, response = "Ozone")
myDf <- as.data.frame(myMat)
myDf

pdpPairs

Description

Creates a pairs plot showing bivariate pdp on upper diagonal, ice/univariate pdp on the diagonal and data on the lower diagonal

Usage

pdpPairs(
  data,
  fit,
  response,
  vars = NULL,
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  fitlims = "pdp",
  gridSize = 10,
  nmax = 500,
  class = 1,
  nIce = 30,
  colorVar = NULL,
  comboImage = FALSE,
  predictFun = NULL,
  convexHull = FALSE,
  probability = FALSE
)
pdpPairs(
  data,
  fit,
  response,
  vars = NULL,
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  fitlims = "pdp",
  gridSize = 10,
  nmax = 500,
  class = 1,
  nIce = 30,
  colorVar = NULL,
  comboImage = FALSE,
  predictFun = NULL,
  convexHull = FALSE,
  probability = FALSE
)

Arguments

`data`	Data frame used for fit.
`fit`	A supervised machine learning model, which understands condvis2::CVpredict
`response`	The name of the response for the fit.
`vars`	The variables to plot (and their order), defaults to all variables other than response.
`pal`	A vector of colors to show predictions, for use with scale_fill_gradientn
`fitlims`	Specifies the fit range for the color map. Options are a numeric vector of length 2, "pdp" (default), in which cases limits are calculated from the pdp, or "all", when limits are calculated from the observations and pdp. Predictions outside fitlims are squished on the color scale.
`gridSize`	The size of the grid for evaluating the predictions.
`nmax`	Uses sample of nmax data rows for the pdp. Default is 500. Use all rows if NULL.
`class`	Category for classification, a factor level, or a number indicating which factor level.
`nIce`	Number of ice curves to be plotted, defaults to 30.
`colorVar`	Which variable to colour the predictions by.
`comboImage`	If TRUE draws pdp for mixed variable plots as an image, otherwise an interaction plot.
`predictFun`	Function of (fit, data) to extract numeric predictions from fit. Uses condvis2::CVpredict by default, which works for many fit classes.
`convexHull`	If TRUE, then the convex hull is computed and any points outside the convex hull are removed.
`probability`	if TRUE, then returns the partial dependence for classification on the probability scale. If FALSE (default), then the partial dependence is returned on a near logit scale.

Value

A pairs plot

Examples

# Load in the data:
aq <- na.omit(airquality)
f <- lm(Ozone ~ ., data = aq)
pdpPairs(aq, f, "Ozone")

# Run a ranger model:
library(ranger)
library(MASS)
Boston1 <- Boston[, c(4:6, 8, 13:14)]
Boston1$chas <- factor(Boston1$chas)
fit <- ranger(medv ~ ., data = Boston1, importance = "permutation")
pdpPairs(Boston1[1:30, ], fit, "medv")
pdpPairs(Boston1[1:30, ], fit, "medv", comboImage = TRUE)
viv <- vivi(Boston1, fit, "medv")
# show top variables only
pdpPairs(Boston1[1:30, ], fit, "medv", comboImage = TRUE, vars = rownames(viv)[1:4])


library(ranger)
rf <- ranger(Species ~ ., data = iris, probability = TRUE)
pdpPairs(iris, rf, "Species") # prediction probs for first class, setosa
pdpPairs(iris, rf, "Species", class = "versicolor") # prediction probs versicolor

# Load in the data:
aq <- na.omit(airquality)
f <- lm(Ozone ~ ., data = aq)
pdpPairs(aq, f, "Ozone")

# Run a ranger model:
library(ranger)
library(MASS)
Boston1 <- Boston[, c(4:6, 8, 13:14)]
Boston1$chas <- factor(Boston1$chas)
fit <- ranger(medv ~ ., data = Boston1, importance = "permutation")
pdpPairs(Boston1[1:30, ], fit, "medv")
pdpPairs(Boston1[1:30, ], fit, "medv", comboImage = TRUE)
viv <- vivi(Boston1, fit, "medv")
# show top variables only
pdpPairs(Boston1[1:30, ], fit, "medv", comboImage = TRUE, vars = rownames(viv)[1:4])


library(ranger)
rf <- ranger(Species ~ ., data = iris, probability = TRUE)
pdpPairs(iris, rf, "Species") # prediction probs for first class, setosa
pdpPairs(iris, rf, "Species", class = "versicolor") # prediction probs versicolor

pdpVars

Description

Displays the individual conditional expectation (ICE) curves and aggregated partial dependence for each variable in a grid.

Usage

pdpVars(
  data,
  fit,
  response,
  vars = NULL,
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  gridSize = 10,
  nmax = 500,
  class = 1,
  nIce = 30,
  predictFun = NULL,
  limits = NULL,
  colorVar = NULL,
  draw = TRUE,
  probability = FALSE
)
pdpVars(
  data,
  fit,
  response,
  vars = NULL,
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  gridSize = 10,
  nmax = 500,
  class = 1,
  nIce = 30,
  predictFun = NULL,
  limits = NULL,
  colorVar = NULL,
  draw = TRUE,
  probability = FALSE
)

Arguments

`data`	Data frame used for fit.
`fit`	A supervised machine learning model, which understands condvis2::CVpredict
`response`	The name of the response for the fit.
`vars`	The variables to plot (and their order), defaults to all variables other than response.
`pal`	A vector of colors to show predictions, for use with scale_fill_gradientn
`gridSize`	The size of the grid for evaluating the predictions.
`nmax`	Uses sample of nmax data rows for the pdp. Default is 500. Use all rows if NULL.
`class`	Category for classification, a factor level, or a number indicating which factor level.
`nIce`	Number of ice curves to be plotted, defaults to 30.
`predictFun`	Function of (fit, data) to extract numeric predictions from fit. Uses condvis2::CVpredict by default, which works for many fit classes.
`limits`	A vector determining the limits of the predicted values.
`colorVar`	Which variable to colour the predictions by.
`draw`	If FALSE, then the plot will not be drawn. Default is TRUE.
`probability`	if TRUE, then returns the partial dependence for classification on the probability scale. If FALSE (default), then the partial dependence is returned on a near logit scale.

Value

A grid displaying ICE curves and univariate partial dependence.

Examples


# Load in the data:
aq <- na.omit(airquality)
fit <- lm(Ozone ~ ., data = aq)
pdpVars(aq, fit, "Ozone")

# Classification
library(ranger)
rfClassif <- ranger(Species ~ ., data = iris, probability = TRUE)
pdpVars(iris, rfClassif, "Species", class = 3)

pp <- pdpVars(iris, rfClassif, "Species", class = 2, draw = FALSE)
pp[[1]]
pdpVars(iris, rfClassif, "Species", class = 2, colorVar = "Species")

# Load in the data:
aq <- na.omit(airquality)
fit <- lm(Ozone ~ ., data = aq)
pdpVars(aq, fit, "Ozone")

# Classification
library(ranger)
rfClassif <- ranger(Species ~ ., data = iris, probability = TRUE)
pdpVars(iris, rfClassif, "Species", class = 3)

pp <- pdpVars(iris, rfClassif, "Species", class = 2, draw = FALSE)
pp[[1]]
pdpVars(iris, rfClassif, "Species", class = 2, colorVar = "Species")

Create a zenplot displaying partial dependence values.

Description

Constructs a zigzag expanded navigation plot (zenplot) displaying partial dependence values.

Usage

pdpZen(
  data,
  fit,
  response,
  zpath = NULL,
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  fitlims = "pdp",
  gridSize = 10,
  nmax = 500,
  class = 1,
  comboImage = FALSE,
  rug = TRUE,
  predictFun = NULL,
  convexHull = FALSE,
  probability = FALSE,
  ...
)
pdpZen(
  data,
  fit,
  response,
  zpath = NULL,
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  fitlims = "pdp",
  gridSize = 10,
  nmax = 500,
  class = 1,
  comboImage = FALSE,
  rug = TRUE,
  predictFun = NULL,
  convexHull = FALSE,
  probability = FALSE,
  ...
)

Arguments

`data`	Data frame used for fit
`fit`	A supervised machine learning model, which understands condvis2::CVpredict
`response`	The name of the response for the fit
`zpath`	Plot shows consecutive pairs of these variables. Defaults to all variables other than response. Recommend constructing zpath witn `calcZpath`.
`pal`	A vector of colors to show predictions, for use with scale_fill_gradientn
`fitlims`	Specifies the fit range for the color map. Options are a numeric vector of length 2, "pdp" (default), in which cases limits are calculated from the pdp, or "all", when limits are calculated from the observations and pdp predictions outside fitlims are squished on the color scale.
`gridSize`	The size of the grid for evaluating the predictions.
`nmax`	Uses sample of nmax data rows for the pdp. Default is 500. Use all rows if NULL.
`class`	Category for classification, a factor level, or a number indicating which factor level.
`comboImage`	If TRUE draws pdp for mixed variable plots as an image, otherwise an interaction plot.
`rug`	If TRUE adds rugs for the data to the pdp plots
`predictFun`	Function of (fit, data) to extract numeric predictions from fit. Uses condvis2::CVpredict by default, which works for many fit classes.
`convexHull`	If TRUE, then the convex hull is computed and any points outside the convex hull are removed.
`probability`	if TRUE, then returns the partial dependence for classification on the probability scale. If FALSE (default), then the partial dependence is returned on a near logit scale.
`...`	passed on to zenplot

Value

A zenplot of partial dependence values.

Examples

## Not run: 
# To use this function, install zenplots and graph from Bioconductor.
if (!requireNamespace("graph", quietly = TRUE)) {
  install.packages("BiocManager")
  BiocManager::install("graph")
}
install.packages("zenplots")

library(MASS)
library(ranger)
Boston1 <- Boston
Boston1$chas <- factor(Boston1$chas)
rf <- ranger(medv ~ ., data = Boston1)
pdpZen(Boston1[1:30, ], rf, response = "medv", zpath = names(Boston1)[1:4], comboImage = T)
# Find the top variables in rf
set.seed(123)
viv <- vivi(Boston1, rf, "medv", nmax = 30) # use 30 rows, for speed
pdpZen(Boston1, rf, response = "medv", zpath = rownames(viv)[1:4], comboImage = T)
zpath <- zPath(viv, cutoff = .2) # find plots whose interaction score exceeds .2
pdpZen(Boston1, rf, response = "medv", zpath = zpath, comboImage = T)

## End(Not run)
## Not run: 
# To use this function, install zenplots and graph from Bioconductor.
if (!requireNamespace("graph", quietly = TRUE)) {
  install.packages("BiocManager")
  BiocManager::install("graph")
}
install.packages("zenplots")

library(MASS)
library(ranger)
Boston1 <- Boston
Boston1$chas <- factor(Boston1$chas)
rf <- ranger(medv ~ ., data = Boston1)
pdpZen(Boston1[1:30, ], rf, response = "medv", zpath = names(Boston1)[1:4], comboImage = T)
# Find the top variables in rf
set.seed(123)
viv <- vivi(Boston1, rf, "medv", nmax = 30) # use 30 rows, for speed
pdpZen(Boston1, rf, response = "medv", zpath = rownames(viv)[1:4], comboImage = T)
zpath <- zPath(viv, cutoff = .2) # find plots whose interaction score exceeds .2
pdpZen(Boston1, rf, response = "medv", zpath = zpath, comboImage = T)

## End(Not run)

vip2vivid

Description

Takes measured importance and interactions from the vip package and turns them into a matrix which can be used for plotting. Accepts any of the variable importance methods supplied by vip.

Usage

vip2vivid(importance, interaction, reorder = TRUE)
vip2vivid(importance, interaction, reorder = TRUE)

Arguments

`importance`	Measured importance from the vip package using `vi` function.
`interaction`	Measured interaction from the vip package using `vint` function.
`reorder`	If TRUE (default) uses DendSer to reorder the matrix of interactions and variable importances.

Value

A matrix of interaction values, with importance on the diagonal.

Examples

## Not run: 
library(ranger)
library(vip)
aq <- na.omit(airquality) # get data
nameAq <- names(aq[-1]) # get feature names

rF <- ranger(Ozone ~ ., data = aq, importance = "permutation") # create ranger random forest fit
vImp <- vi(rF) # vip importance
vInt <- vint(rF, feature_names = nameAq) # vip interaction

vip2vivid(vImp, vInt)

## End(Not run)
## Not run: 
library(ranger)
library(vip)
aq <- na.omit(airquality) # get data
nameAq <- names(aq[-1]) # get feature names

rF <- ranger(Ozone ~ ., data = aq, importance = "permutation") # create ranger random forest fit
vImp <- vi(rF) # vip importance
vInt <- vint(rF, feature_names = nameAq) # vip interaction

vip2vivid(vImp, vInt)

## End(Not run)

vivi

Description

Creates a matrix displaying variable importance on the diagonal and variable interaction on the off-diagonal.

Usage

vivi(
  data,
  fit,
  response,
  gridSize = 50,
  importanceType = "agnostic",
  nmax = 500,
  reorder = TRUE,
  class = 1,
  predictFun = NULL,
  normalized = FALSE,
  numPerm = 4,
  showVimpError = FALSE,
  vars = NULL
)
vivi(
  data,
  fit,
  response,
  gridSize = 50,
  importanceType = "agnostic",
  nmax = 500,
  reorder = TRUE,
  class = 1,
  predictFun = NULL,
  normalized = FALSE,
  numPerm = 4,
  showVimpError = FALSE,
  vars = NULL
)

Arguments

`data`	Data frame used for fit.
`fit`	A supervised machine learning model, which understands condvis2::CVpredict
`response`	The name of the response for the fit.
`gridSize`	The size of the grid for evaluating the predictions.
`importanceType`	Used to select the importance metric. By default, an agnostic importance measure is used. If an embedded metric is available, then setting this argument to the importance metric will use the selected importance values in the vivid-matrix. Please refer to the examples given for illustration. Alternatively, set to equal "agnostic" (the default) to override embedded importance measures and return agnostic importance values.
`nmax`	Maximum number of data rows to consider. Default is 500. Use all rows if NULL.
`reorder`	If TRUE (default) uses DendSer to reorder the matrix of interactions and variable importances.
`class`	Category for classification, a factor level, or a number indicating which factor level.
`predictFun`	Function of (fit, data) to extract numeric predictions from fit. Uses condvis2::CVpredict by default, which works for many fit classes.
`normalized`	Should Friedman's H-statistic be normalized or not. Default is FALSE.
`numPerm`	Number of permutations to perform for agnostic importance. Default is 4.
`showVimpError`	Logical. If TRUE, and `numPerm > 1` then a tibble containing the variable names, their importance values, and the standard error for each importance is printed to the console.
`vars`	A vector of variable names to be assessed.

Details

If the argument importanceType = 'agnostic', then an agnostic permutation importance (1) is calculated. Friedman's H statistic (2) is used for measuring the interactions. This measure is based on partial dependence curves and relates the interaction strength of a pair of variables to the total effect strength of that variable pair.

Value

A matrix of interaction values, with importance on the diagonal.

References

1: Fisher A., Rudin C., Dominici F. (2018). All Models are Wrong but many are Useful: Variable Importance for Black-Box, Proprietary, or Misspecified Prediction Models, using Model Class Reliance. Arxiv.

2: Friedman, J. H. and Popescu, B. E. (2008). “Predictive learning via rule ensembles.” The Annals of Applied Statistics. JSTOR, 916–54.

Examples


aq <- na.omit(airquality)
f <- lm(Ozone ~ ., data = aq)
m <- vivi(fit = f, data = aq, response = "Ozone") # as expected all interactions are zero
viviHeatmap(m)

# Select importance metric
library(randomForest)
rf1 <- randomForest(Ozone~., data = aq, importance = TRUE)
m2 <- vivi(fit = rf1, data = aq, response = 'Ozone',
           importanceType = '%IncMSE') # select %IncMSE as the importance measure
viviHeatmap(m2)


library(ranger)
rf <- ranger(Species ~ ., data = iris, importance = "impurity", probability = TRUE)
vivi(fit = rf, data = iris, response = "Species") # returns agnostic importance
vivi(fit = rf, data = iris, response = "Species",
     importanceType = "impurity") # returns selected 'impurity' importance.

aq <- na.omit(airquality)
f <- lm(Ozone ~ ., data = aq)
m <- vivi(fit = f, data = aq, response = "Ozone") # as expected all interactions are zero
viviHeatmap(m)

# Select importance metric
library(randomForest)
rf1 <- randomForest(Ozone~., data = aq, importance = TRUE)
m2 <- vivi(fit = rf1, data = aq, response = 'Ozone',
           importanceType = '%IncMSE') # select %IncMSE as the importance measure
viviHeatmap(m2)


library(ranger)
rf <- ranger(Species ~ ., data = iris, importance = "impurity", probability = TRUE)
vivi(fit = rf, data = iris, response = "Species") # returns agnostic importance
vivi(fit = rf, data = iris, response = "Species",
     importanceType = "impurity") # returns selected 'impurity' importance.

vividReorder

Description

Reorders a square matrix so that values of high importance and interaction strength are pushed to the top left of the matrix.

Usage

vividReorder(d)
vividReorder(d)

Arguments

`d`	A matrix such as that returned by vivi

Value

A reordered version of d.

Examples

f <- lm(Sepal.Length ~ ., data = iris[, -5])
m <- vivi(fit = f, data = iris[, -5], response = "Sepal.Length")
corimp <- abs(cor(iris[, -5])[1, -1])
viviUpdate(m, corimp) # use correlation as importance and reorder

f <- lm(Sepal.Length ~ ., data = iris[, -5])
m <- vivi(fit = f, data = iris[, -5], response = "Sepal.Length")
corimp <- abs(cor(iris[, -5])[1, -1])
viviUpdate(m, corimp) # use correlation as importance and reorder

viviHeatmap

Description

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

Usage

viviHeatmap(
  mat,
  intPal = rev(colorspace::sequential_hcl(palette = "Purples 3", n = 100)),
  impPal = rev(colorspace::sequential_hcl(palette = "Greens 3", n = 100)),
  intLims = NULL,
  impLims = NULL,
  border = FALSE,
  angle = 0
)
viviHeatmap(
  mat,
  intPal = rev(colorspace::sequential_hcl(palette = "Purples 3", n = 100)),
  impPal = rev(colorspace::sequential_hcl(palette = "Greens 3", n = 100)),
  intLims = NULL,
  impLims = NULL,
  border = FALSE,
  angle = 0
)

Arguments

`mat`	A matrix, such as that returned by vivi, of values to be plotted.
`intPal`	A vector of colours to show interactions, for use with scale_fill_gradientn.
`impPal`	A vector of colours to show importance, for use with scale_fill_gradientn.
`intLims`	Specifies the fit range for the color map for interaction strength.
`impLims`	Specifies the fit range for the color map for importance.
`border`	Logical. If TRUE then draw a black border around the diagonal elements.
`angle`	The angle to rotate the x-axis labels. Defaults to zero.

Value

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

Examples


library(ranger)
aq <- na.omit(airquality)
rF <- ranger(Ozone ~ ., data = aq, importance = "permutation")
myMat <- vivi(fit = rF, data = aq, response = "Ozone")
viviHeatmap(myMat)

library(ranger)
aq <- na.omit(airquality)
rF <- ranger(Ozone ~ ., data = aq, importance = "permutation")
myMat <- vivi(fit = rF, data = aq, response = "Ozone")
viviHeatmap(myMat)

viviNetwork

Description

Create a Network plot displaying variable importance and variable interaction.

Usage

viviNetwork(
  mat,
  intThreshold = NULL,
  intLims = NULL,
  impLims = NULL,
  intPal = rev(colorspace::sequential_hcl(palette = "Purples 3", n = 100)),
  impPal = rev(colorspace::sequential_hcl(palette = "Greens 3", n = 100)),
  removeNode = FALSE,
  layout = igraph::layout_in_circle,
  cluster = NULL,
  nudge_x = 0.05,
  nudge_y = 0.03,
  edgeWidths = 1:4
)
viviNetwork(
  mat,
  intThreshold = NULL,
  intLims = NULL,
  impLims = NULL,
  intPal = rev(colorspace::sequential_hcl(palette = "Purples 3", n = 100)),
  impPal = rev(colorspace::sequential_hcl(palette = "Greens 3", n = 100)),
  removeNode = FALSE,
  layout = igraph::layout_in_circle,
  cluster = NULL,
  nudge_x = 0.05,
  nudge_y = 0.03,
  edgeWidths = 1:4
)

Arguments

`mat`	A matrix, such as that returned by vivi, of values to be plotted.
`intThreshold`	Remove edges with weight below this value if provided.
`intLims`	Specifies the fit range for the color map for interaction strength.
`impLims`	Specifies the fit range for the color map for importance.
`intPal`	A vector of colours to show interactions, for use with scale_fill_gradientn.
`impPal`	A vector of colours to show importance, for use with scale_fill_gradientn.
`removeNode`	If TRUE, then removes nodes with no connecting edges when thresholding interaction values.
`layout`	igraph layout function or a numeric matrix with two columns, one row per node. Defaults to igraph::layout_as_circle
`cluster`	Either a vector of cluster memberships for nodes or an igraph clustering function.
`nudge_x`	Nudge (centered) labels by this amount, outward horizontally.
`nudge_y`	Nudge (centered) labels by this amount, outward vertically.
`edgeWidths`	A vector specifying the scaling of the edges for the displayed graph. Values must be positive.

Value

A plot displaying interaction strength between variables on the edges and variable importance on the nodes.

Examples


library(ranger)
aq <- na.omit(airquality)
rF <- ranger(Ozone ~ ., data = aq, importance = "permutation")
myMat <- vivi(fit = rF, data = aq, response = "Ozone")
viviNetwork(myMat)

library(ranger)
aq <- na.omit(airquality)
rF <- ranger(Ozone ~ ., data = aq, importance = "permutation")
myMat <- vivi(fit = rF, data = aq, response = "Ozone")
viviNetwork(myMat)

viviUpdate

Description

Creates a matrix displaying updated variable importance on the diagonal and variable interaction on the off-diagonal.

Usage

viviUpdate(mat, newImp, reorder = TRUE)
viviUpdate(mat, newImp, reorder = TRUE)

Arguments

`mat`	A matrix, such as that returned by `vivi`.
`newImp`	A named vector of variable importances.
`reorder`	If TRUE (default) uses DendSer to reorder the matrix of interactions and variable importances.

Value

A matrix of values, of class vivid, with updated variable importances.

Examples

f <- lm(Sepal.Length ~ ., data = iris[, -5])
m <- vivi(iris[, -5], f, "Sepal.Length")
corimp <- abs(cor(iris[, -5])[1, -1])
viviUpdate(m, corimp) # use correlation as updated importance
f <- lm(Sepal.Length ~ ., data = iris[, -5])
m <- vivi(iris[, -5], f, "Sepal.Length")
corimp <- abs(cor(iris[, -5])[1, -1])
viviUpdate(m, corimp) # use correlation as updated importance

zPath

Description

Constructs a zenpath for connecting and displaying pairs.

Usage

zPath(
  viv,
  cutoff = NULL,
  method = c("greedy.weighted", "strictly.weighted"),
  connect = TRUE
)
zPath(
  viv,
  cutoff = NULL,
  method = c("greedy.weighted", "strictly.weighted"),
  connect = TRUE
)

Arguments

`viv`	A matrix, created by `vivi` to be used to calculate the path.
`cutoff`	Do not include any variables that are below the cutoff interaction value.
`method`	String indicating the method to use. The available methods are: "greedy.weighted": Sort all pairs according to a greedy (heuristic) Euler path with x as weights visiting each edge precisely once. "strictly.weighted": Strictly respect the order of the weights - so the first, second, third, and so on, adjacent pair of numbers of the output of zenpath() corresponds to the pair with largest, second-largest, third-largest, and so on, weight. see zenpath
`connect`	If connect is TRUE, connect the edges from separate eulerians (strictly.weighted only).

Details

Construct a path of indices to visit to order variables

Value

Returns a zpath from viv showing pairs with viv entry over the cutoff

Examples

## Not run: 
# To use this function, install zenplots and graph from Bioconductor.
if (!requireNamespace("graph", quietly = TRUE)) {
  install.packages("BiocManager")
  BiocManager::install("graph")
}
install.packages("zenplots")

aq <- na.omit(airquality) * 1.0

# Run an mlr3 ranger model:
library(mlr3)
library(mlr3learners)
library(ranger)
ozonet <- TaskRegr$new(id = "airQ", backend = aq, target = "Ozone")
ozonel <- lrn("regr.ranger", importance = "permutation")
ozonef <- ozonel$train(ozonet)

viv <- vivi(aq, ozonef, "Ozone")

# Calculate Zpath:
zpath <- zPath(viv, .8)
zpath

## End(Not run)
## Not run: 
# To use this function, install zenplots and graph from Bioconductor.
if (!requireNamespace("graph", quietly = TRUE)) {
  install.packages("BiocManager")
  BiocManager::install("graph")
}
install.packages("zenplots")

aq <- na.omit(airquality) * 1.0

# Run an mlr3 ranger model:
library(mlr3)
library(mlr3learners)
library(ranger)
ozonet <- TaskRegr$new(id = "airQ", backend = aq, target = "Ozone")
ozonel <- lrn("regr.ranger", importance = "permutation")
ozonef <- ozonel$train(ozonet)

viv <- vivi(aq, ozonef, "Ozone")

# Calculate Zpath:
zpath <- zPath(viv, .8)
zpath

## End(Not run)

Package 'vivid'

Help Index

as.data.frame.vivid

Description

Usage

Arguments

Value

Examples

pdpPairs

Description

Usage

Arguments

Value

Examples

pdpVars

Description

Usage

Arguments

Value

Examples

Create a zenplot displaying partial dependence values.

Description

Usage

Arguments

Value

Examples

vip2vivid

Description

Usage

Arguments

Value

Examples

vivi

Description

Usage

Arguments

Details

Value

References

Examples

vividReorder

Description

Usage

Arguments

Value

Examples

viviHeatmap

Description

Usage

Arguments

Value

Examples

viviNetwork

Description

Usage

Arguments

Value

Examples

viviUpdate

Description

Usage

Arguments

Value

Examples

zPath

Description

Usage

Arguments

Details

Value

Examples