predict_PDP.2D.Plotter {distillML}R Documentation

Two Dimensional Prediction Curve for PDP Plots


Gives prediction surface for all specified feature pairs in the interpreter object (features.2d)


predict_PDP.2D.Plotter(object, feat.2d, save = TRUE)



The Interpreter object to use.


A 2-column dataframe or matrix that gives the first variable in in the first column, and the second variable in the next. The number of rows is equal to the number of 2-D PDPs one would like.


A boolean indicator to indicate whether the calculations should be saved in the interpreter object or not. This can help reduce computation if the PDP functions are used many times, but requires additional memory to store the predictions. By default, this is set to TRUE.




A list of data frames for each pair of features.2d. Each data frame contains columns corresponding to the grid points for the two selected features and a column corresponding to the predictions of the model at the given combination of grid points.


This method is meant to primarily be used to find the exact values for the 2-D PDP curves or heatmap plotted. Note that after the PDP curve is plotted, the returned object of this function will be the saved predictions for plotting the curve, rather than a recalculation of the values.

[Package distillML version Index]