model_metrics: Model Metrics and Performance

In lares: Analytics & Machine Learning Sidekick

Model Metrics and Performance

Description

This function lets the user get a confusion matrix and accuracy, and for for binary classification models: AUC, Precision, Sensitivity, and Specificity, given the expected (tags) values and predicted values (scores).

Usage

model_metrics(
  tag,
  score,
  multis = NA,
  abc = TRUE,
  thresh = 10,
  auto_n = TRUE,
  thresh_cm = 0.5,
  target = "auto",
  type = "test",
  model_name = NA,
  plots = TRUE,
  quiet = FALSE,
  subtitle = NA
)

Arguments

tag	Vector. Real known label
score	Vector. Predicted value or model's result
multis	Data.frame. Containing columns with each category score (only used when more than 2 categories coexist)
abc	Boolean. Arrange columns and rows alphabetically when categorical values?
thresh	Integer. Threshold for selecting binary or regression models: this number is the threshold of unique values we should have in 'tag' (more than: regression; less than: classification)
auto_n	Add n_ before digits when it's categorical and not numerical, even though seems numerical?
thresh_cm	Numeric. Value to splits the results for the confusion matrix. Range of values: (0-1)
target	Value. Which is your target positive value? If set to 'auto', the target with largest mean(score) will be selected. Change the value to overwrite. Only used when binary categorical model.
type	Character. One of: "train", "test".
model_name	Character. Model's name for reference.
plots	Boolean. Create plots objects?
quiet	Boolean. Quiet all messages, warnings, recommendations?
subtitle	Character. Subtitle for plots

Value

List. Multiple performance metrics that vary depending on the type of model (classification or regression). If plot=TRUE, multiple plots are also returned.

See Also

Other Machine Learning: ROC(), conf_mat(), export_results(), gain_lift(), h2o_automl(), h2o_predict_MOJO(), h2o_selectmodel(), impute(), iter_seeds(), lasso_vars(), model_preprocess(), msplit()

Other Model metrics: ROC(), conf_mat(), errors(), gain_lift(), loglossBinary()

Other Calculus: corr(), dist2d(), quants()

Examples

data(dfr) # Results for AutoML Predictions
lapply(dfr, head)

# Metrics for Binomial Model
met1 <- model_metrics(dfr$class2$tag, dfr$class2$scores,
  model_name = "Titanic Survived Model",
  plots = FALSE
)
print(met1)

# Metrics for Multi-Categorical Model
met2 <- model_metrics(dfr$class3$tag, dfr$class3$score,
  multis = subset(dfr$class3, select = -c(tag, score)),
  model_name = "Titanic Class Model",
  plots = FALSE
)
print(met2)

# Metrics for Regression Model
met3 <- model_metrics(dfr$regr$tag, dfr$regr$score,
  model_name = "Titanic Fare Model",
  plots = FALSE
)
print(met3)

lares documentation built on Sept. 13, 2024, 1:08 a.m.