Random Hazard Forests

Random Hazard Forests (RHF) extend Random Survival Forests (RSF) by directly estimating the hazard function and by accommodating time-dependent covariates through counting-process style inputs. The package fits tree ensembles for dynamic survival prediction, returning hazard, cumulative hazard, integrated hazard, and related performance summaries for training and test data. The methods build on Random Survival Forests described by Ishwaran et al. (2008) and on nonparametric hazard modeling with time-dependent covariates described by Lee et al. (2021) .

A Modified Random Survival Forest Algorithm

Implements a modification to the Random Survival Forests algorithm for obtaining variable importance in high dimensional datasets. The proposed algorithm is appropriate for settings in which a silent event is observed through sequentially administered, error-prone self-reports or laboratory based diagnostic tests. The modified algorithm incorporates a formal likelihood framework that accommodates sequentially administered, error-prone self-reports or laboratory based diagnostic tests. The original Random Survival Forests algorithm is modified by the introduction of a new splitting criterion based on a likelihood ratio test statistic.

Random Survival Forest for Recurrent Events

A tool designed to analyze recurrent events when dealing with right-censored data and the potential presence of a terminal event (that prevents further occurrences, like death). It extends the random survival forest algorithm, adapting splitting rules and node estimators to handle complexities of recurrent events. The methodology is fully described in Murris, J., Bouaziz, O., Jakubczak, M., Katsahian, S., & Lavenu, A. (2024) (< https://hal.science/hal-04612431v1/document>).

Classification, Regression and Feature Evaluation

A suite of machine learning algorithms written in C++ with the R interface contains several learning techniques for classification and regression. Predictive models include e.g., classification and regression trees with optional constructive induction and models in the leaves, random forests, kNN, naive Bayes, and locally weighted regression. All predictions obtained with these models can be explained and visualized with the 'ExplainPrediction' package. This package is especially strong in feature evaluation where it contains several variants of Relief algorithm and many impurity based attribute evaluation functions, e.g., Gini, information gain, MDL, and DKM. These methods can be used for feature selection or discretization of numeric attributes. The OrdEval algorithm and its visualization is used for evaluation of data sets with ordinal features and class, enabling analysis according to the Kano model of customer satisfaction. Several algorithms support parallel multithreaded execution via OpenMP. The top-level documentation is reachable through ?CORElearn.

RF Variable Importance for Arbitrary Measures

Computes the random forest variable importance (VIMP) for the conditional inference random forest (cforest) of the 'party' package. Includes a function (varImp) that computes the VIMP for arbitrary measures from the 'measures' package. For calculating the VIMP regarding the measures accuracy and AUC two extra functions exist (varImpACC and varImpAUC).

Random Uniform Forests for Classification, Regression and Unsupervised Learning

Ensemble model, for classification, regression and unsupervised learning, based on a forest of unpruned and randomized binary decision trees. Each tree is grown by sampling, with replacement, a set of variables at each node. Each cut-point is generated randomly, according to the continuous Uniform distribution. For each tree, data are either bootstrapped or subsampled. The unsupervised mode introduces clustering, dimension reduction and variable importance, using a three-layer engine. Random Uniform Forests are mainly aimed to lower correlation between trees (or trees residuals), to provide a deep analysis of variable importance and to allow native distributed and incremental learning.

Fuzzy Forests

Fuzzy forests, a new algorithm based on random forests, is designed to reduce the bias seen in random forest feature selection caused by the presence of correlated features. Fuzzy forests uses recursive feature elimination random forests to select features from separate blocks of correlated features where the correlation within each block of features is high and the correlation between blocks of features is low. One final random forest is fit using the surviving features. This package fits random forests using the 'randomForest' package and allows for easy use of 'WGCNA' to split features into distinct blocks. See D. Conn, Ngun, T., C. Ramirez, and G. Li (2019) for further details.

Quantile Regression Forests for 'ranger'

This is the implementation of quantile regression forests for the fast random forest package 'ranger'.

Recursive Partitioning for Modeling Survey Data

Functions to allow users to build and analyze design consistent tree and random forest models using survey data from a complex sample design. The tree model algorithm can fit a linear model to survey data in each node obtained by recursively partitioning the data. The splitting variables and selected splits are obtained using a randomized permutation test procedure which adjusted for complex sample design features used to obtain the data. Likewise the model fitting algorithm produces design-consistent coefficients to any specified least squares linear model between the dependent and independent variables used in the end nodes. The main functions return the resulting binary tree or random forest as an object of "rpms" or "rpms_forest" type. The package also provides methods modeling a "boosted" tree or forest model and a tree model for zero-inflated data as well as a number of functions and methods available for use with these object types.

Nearest Neighbor Observation Imputation and Evaluation Tools

Performs nearest neighbor-based imputation using one or more alternative approaches to processing multivariate data. These include methods based on canonical correlation: analysis, canonical correspondence analysis, and a multivariate adaptation of the random forest classification and regression techniques of Leo Breiman and Adele Cutler. Additional methods are also offered. The package includes functions for comparing the results from running alternative techniques, detecting imputation targets that are notably distant from reference observations, detecting and correcting for bias, bootstrapping and building ensemble imputations, and mapping results.