xgboost: xgb.plot.importance – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

xgb.plot.importance

Plot feature importance as a bar graph

Description

Represents previously calculated feature importance as a bar graph. xgb.plot.importance uses base R graphics, while xgb.ggplot.importance uses the ggplot backend.

Usage

xgb.ggplot.importance(
  importance_matrix = NULL,
  top_n = NULL,
  measure = NULL,
  rel_to_first = FALSE,
  n_clusters = c(1:10),
  ...
)

xgb.plot.importance(
  importance_matrix = NULL,
  top_n = NULL,
  measure = NULL,
  rel_to_first = FALSE,
  left_margin = 10,
  cex = NULL,
  plot = TRUE,
  ...
)

Arguments

`importance_matrix`	a `data.table` returned by `xgb.importance`.
`top_n`	maximal number of top features to include into the plot.
`measure`	the name of importance measure to plot. When `NULL`, 'Gain' would be used for trees and 'Weight' would be used for gblinear.
`rel_to_first`	whether importance values should be represented as relative to the highest ranked feature. See Details.
`n_clusters`	(ggplot only) a `numeric` vector containing the min and the max range of the possible number of clusters of bars.
`...`	other parameters passed to `barplot` (except horiz, border, cex.names, names.arg, and las).
`left_margin`	(base R barplot) allows to adjust the left margin size to fit feature names. When it is NULL, the existing `par('mar')` is used.
`cex`	(base R barplot) passed as `cex.names` parameter to `barplot`.
`plot`	(base R barplot) whether a barplot should be produced. If FALSE, only a data.table is returned.

Details

The graph represents each feature as a horizontal bar of length proportional to the importance of a feature. Features are shown ranked in a decreasing importance order. It works for importances from both gblinear and gbtree models.

When rel_to_first = FALSE, the values would be plotted as they were in importance_matrix. For gbtree model, that would mean being normalized to the total of 1 ("what is feature's importance contribution relative to the whole model?"). For linear models, rel_to_first = FALSE would show actual values of the coefficients. Setting rel_to_first = TRUE allows to see the picture from the perspective of "what is feature's importance contribution relative to the most important feature?"

The ggplot-backend method also performs 1-D clustering of the importance values, with bar colors corresponding to different clusters that have somewhat similar importance values.

Value

The xgb.plot.importance function creates a barplot (when plot=TRUE) and silently returns a processed data.table with n_top features sorted by importance.

The xgb.ggplot.importance function returns a ggplot graph which could be customized afterwards. E.g., to change the title of the graph, add + ggtitle("A GRAPH NAME") to the result.

Examples

data(agaricus.train)

bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 3,
               eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")

importance_matrix <- xgb.importance(colnames(agaricus.train$data), model = bst)

xgb.plot.importance(importance_matrix, rel_to_first = TRUE, xlab = "Relative importance")

(gg <- xgb.ggplot.importance(importance_matrix, measure = "Frequency", rel_to_first = TRUE))
gg + ggplot2::ylab("Frequency")

xgboost

Extreme Gradient Boosting

v1.4.1.1

Apache License (== 2.0) | file LICENSE

Authors

Tianqi Chen [aut], Tong He [aut, cre], Michael Benesty [aut], Vadim Khotilovich [aut], Yuan Tang [aut] (<https://orcid.org/0000-0001-5243-233X>), Hyunsu Cho [aut], Kailong Chen [aut], Rory Mitchell [aut], Ignacio Cano [aut], Tianyi Zhou [aut], Mu Li [aut], Junyuan Xie [aut], Min Lin [aut], Yifeng Geng [aut], Yutian Li [aut], XGBoost contributors [cph] (base XGBoost implementation)

Initial release

2021-04-22