Innovative Complex Split Procedures in Random Forests Through Candidate Split Sampling
Diversity Forests
Construct a basic diversity forest prediction rule that uses univariab...
Diversity Forest variable importance
Construct an interaction forest prediction rule and calculate EIM valu...
Construct a multi forest prediction rule and calculate multi-class and...
Plot method for interactionfor
objects
Plot method for multifor
objects
Interaction forest plots: exploring interaction forest results through...
Plots of the (estimated) within-class distributions of variables
Plot of the (estimated) simultaneous influence of two variables
Plot of the (estimated) dependency structure of a variable x
on a ca...
Diversity Forest prediction
Interaction Forest prediction
Multi forest prediction
Diversity Forest predictions
Diversity Forest predictions
Optimization of the values of the tuning parameters nsplits
and `pro...
Implementations of three diversity forest (DF) (Hornung, 2022, <doi:10.1007/s42979-021-00920-1>) variants. The DF algorithm is a split-finding approach that allows complex split procedures to be realized in random forest variants. The three DF variants implemented are: 1. interaction forests (IFs) (Hornung & Boulesteix, 2022, <doi:10.1016/j.csda.2022.107460>): Model quantitative and qualitative interaction effects using bivariable splitting. Come with the Effect Importance Measure (EIM), which can be used to identify variable pairs that have well-interpretable quantitative and qualitative interaction effects with high predictive relevance. 2. multi forests (MuFs) (Hornung & Hapfelmeier, 2024, <doi:10.48550/arXiv.2409.08925>): Model multi-class outcomes using multi-way and binary splitting. Come with two variable importance measures (VIMs): The multi-class VIM measures the degree to which the variables are specifically associated with one or more outcome classes, and the discriminatory VIM, similar to conventional VIMs, measures the overall influence strength of the variables. 3. the basic form of diversity forests that uses conventional univariable, binary splitting (Hornung, 2022). Except for multi forests, which are tailored for multi-class outcomes, all included diversity forest variants support categorical, metric, and survival outcomes. The package also includes plotting functions that make it possible to learn about the forms of the effects identified using IFs and MuFs. This is a fork of the R package 'ranger' (main author: Marvin N. Wright), which implements random forests using an efficient C++ implementation.