Partial dependence plot (PDP) gives a graphical depiction of the marginal effect of a variable on the response. The effect of a variable is measured in change in the mean response. PDP assumes independence between the feature for which is the PDP computed and the rest.
h2o.pd_plot( object, newdata, column, target = NULL, row_index = NULL, max_levels = 30, binary_response_scale = c("response", "logodds"), grouping_column = NULL, nbins = 100, show_rug = TRUE )
object | An H2O model. |
---|---|
newdata | An H2OFrame. Used to generate predictions used in Partial Dependence calculations. |
column | A feature column name to inspect. Character string. |
target | If multinomial, plot PDP just for |
row_index | Optional. Calculate Individual Conditional Expectation (ICE) for row, |
max_levels | An integer specifying the maximum number of factor levels to show. Defaults to 30. |
binary_response_scale | Option for binary model to display (on the y-axis) the logodds instead of the actual score. Can be one of: "response", "logodds". Defaults to "response". |
grouping_column | A feature column name to group the data and provide separate sets of plots by grouping feature values |
nbins | A number of bins used. Defaults to 100. |
show_rug | Show rug to visualize the density of the column. Defaults to TRUE. |
A ggplot2 object
# NOT RUN { library(h2o) h2o.init() # Import the wine dataset into H2O: f <- "https://h2o-public-test-data.s3.amazonaws.com/smalldata/wine/winequality-redwhite-no-BOM.csv" df <- h2o.importFile(f) # Set the response response <- "quality" # Split the dataset into a train and test set: splits <- h2o.splitFrame(df, ratios = 0.8, seed = 1) train <- splits[[1]] test <- splits[[2]] # Build and train the model: gbm <- h2o.gbm(y = response, training_frame = train) # Create the partial dependence plot pdp <- h2o.pd_plot(gbm, test, column = "alcohol") print(pdp) # }