Description Usage Arguments Value Figures Author(s) References Examples
Produce a boxplot of performance metrics across resamples for multiple binary classification or regression models from the caret package.
The metrics shown for binary classification are: Area under ROC Curve (AUROC), Sensitivity, Specificity, Area under Precision-Recall Curve (AUPRC), Precision, F1 Score, Accuracy, Cohen's Kappa, Log Loss, Matthews Correlation Coefficient, Concordance, Discordance, Somer's D, KS Statistic, and False Discovery Rate.
The metrics shown for regression are: Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Spearman's Rho, Concordance Correlation Coefficient, and RSquared. Note that MAPE will not be provided if any observations equal zero to avoid division by zero.
Note that the savePredictions argument in caret's trainControl function should be set to "final" so that the resampling results of the model with the optimal tuning parameters are available: trainControl(..., savePredictions = "final")
1 2 3 |
data |
This argument can be either of the below options: 1. A dataframe with the performance metrics of one or more models across resamples. The dataframe must contain the column "Model" containing the model name(s). The remaining columns in the dataframe would be the performance metrics where each metric is one column. In this option, the argument mod_names is not used. 2. A list of caret model objects. In this case, the argument mod_names is required. |
mod_names |
Defaults to NULL. A vector or list of names for the models. This argument is required only if option 2 is used for the data argument. Otherwise, it is not used. |
ncol |
Defaults to NULL. An integer providing the number of facet columns to have in the plot. NULL value would use the default settings of ggplot2's facet_wrap() function. |
nrow |
Defaults to NULL. An integer providing the number of facet rows to have in the plot. NULL value would use the default settings of ggplot2's facet_wrap() function. |
scales |
Defaults to "free_y". A string that is passed to the scales argument of ggplot's facet_wrap() function. |
conf_int95 |
Defaults to FALSE. A logical indicating whether the plot should add errorbars representing the 95% confidence interval of the mean. |
Returns a ggplot object with the box plot of performance metrics of the model objects across resamples.
Andrew Kostandy (andrew.kostandy@gmail.com)
The "InformationValue", "caret", and "MLmetrics" packages were used to compute many of the metrics.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | train_ctrl <- trainControl(method = "repeatedcv",
number = 10,
repeats = 4,
summaryFunction = defaultSummary,
savePredictions = "final")
lm_fit_1 <- train(Sepal.Length ~ Sepal.Width,
data = iris,
method = "lm",
metric = "RMSE",
trControl = train_ctrl)
lm_fit_2 <- train(Sepal.Length ~ Sepal.Width + Petal.Length,
data = iris,
method = "lm",
metric = "RMSE",
trControl = train_ctrl)
plot_mod_results(list(lm_fit_1, lm_fit_2), c("LM 1", "LM 2"))
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.