Get variable importances
get_variable_importance(models, remove_zeros = TRUE, top_n)
model_list object
Remove features with zero variable importance? Default is TRUE
Integer: How many variables to return? The top_n most important variables be returned. If missing (default), all variables are returned
Data frame of variables and their importance for predictive power
Some algorithms provide variable importance, others don't. The best-performing model that offers variable importance will be used.
plot.variable_importance
# NOT RUN { m <- machine_learn(mtcars, outcome = mpg, models = "rf", tune = FALSE) (vi <- get_variable_importance(m)) plot(vi) # }
Run the code above in your browser using DataLab