mplot_splits: Split and compare quantiles plot

In laresbernardo/lares: Analytics & Machine Learning Sidekick

Split and compare quantiles plot

Description

This function lets us split and compare quantiles on a given prediction to compare different categorical values vs scores grouped by equal sized buckets.

Usage

mplot_splits(
  tag,
  score,
  splits = 5,
  subtitle = NA,
  model_name = NA,
  save = FALSE,
  subdir = NA,
  file_name = "viz_splits.png"
)

Arguments

tag	Vector. Real known label.
score	Vector. Predicted value or model's result.
splits	Integer. Number of separations to plot
subtitle	Character. Subtitle to show in plot
model_name	Character. Model's name
save	Boolean. Save output plot into working directory
subdir	Character. Sub directory on which you wish to save the plot
file_name	Character. File name as you wish to save the plot

Value

Plot with distribution and performance results by splits.

See Also

Other ML Visualization: mplot_conf(), mplot_cuts(), mplot_cuts_error(), mplot_density(), mplot_full(), mplot_gain(), mplot_importance(), mplot_lineal(), mplot_metrics(), mplot_response(), mplot_roc(), mplot_topcats()

Examples

Sys.unsetenv("LARES_FONT") # Temporal
data(dfr) # Results for AutoML Predictions
lapply(dfr, head)

# For categorical (binary) values
mplot_splits(dfr$class2$tag, dfr$class2$scores,
  splits = 4,
  model_name = "Titanic Survived Model"
)

# For categorical (+2) values
mplot_splits(dfr$class3$tag, dfr$class2$scores,
  model_name = "Titanic Class Model"
)

# For continuous values
mplot_splits(dfr$regr$tag, dfr$regr$score,
  splits = 4,
  model_name = "Titanic Fare Model"
)

laresbernardo/lares documentation built on Oct. 23, 2024, 12:05 p.m.