An Introduction to Recursive Partitioning: Rationale, Application and Characteristics of Classification and Regression Trees, Bagging and Random Forests

By Carolin Strobl, James D. Malley, Gerhard E. Tutz. Psychological methods, 2009

2,461 citations recorded by the source corpus.

Recursive partitioning methods have become popular and widely used tools for non-parametric regression and classification in many scientific fields. Especially random forests, that can deal with large numbers of predictor variables even in the presence of complex interactions, have been applied successfully in genetics, clinical medicine and bioinformatics within the past few years. High dimensional problems are common not only in genetics, but also in some areas of psychological research, where only few subjects can be measured due to time or cost constraints, yet a large amount of data is generated for each subject. Random forests have been shown to achieve a high prediction accuracy in such applications, and provide descriptive variable importance measures reflecting the impact of each variable in both main effects and interactions. The aim of this work is to introduce the principles of the standard recursive partitioning methods as well as recent methodological improvements, to illustrate their usage for low and high dimensional data exploration, but also to point out limitations of the methods and potential pitfalls in their practical application. Application of the methods is illustrated using freely available implementations in the R system for statistical computing.

Subfields: Artificial Intelligence, Information Systems, Molecular Biology

View Paper: https://www.semanticscholar.org/paper/4544ec96d5ec6e71f7795fa5498f542cb9985c7a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Molecular Biology

Citations in the corpus

The cave of Shadows. Addressing the human factor with generalized additive mixed models (2015, methodology): “Researchers may want to complement exploratory regression analysis with techniques from machine learning such as random forests (Breiman, 2001; Strobl et al., 2009) or gradient boosting machines (Friedman, 2001; Chen et al., 2015) to obtain an independent assessment of variable importance that is…”
How does forest structure affect root reinforcement and susceptibility to shallow landslides? (2016, methodology): “This indicates that they tend to overfit the data (Vorpahl et al., 2012), although random forests are regarded to be resistant against over-fitting (Breiman, 2001; Strobl et al., 2009b).”
Generalized Measurement Invariance Tests with Application to Factor Analysis (2011, methodology): “Such an algorithm is related to classification and regression trees (Breiman, Friedman, Olshen, and Stone 1984; Merkle and Shaffer 2011; Strobl, Malley, and Tutz 2009), with related algorithms being developed for general parametric models (Zeileis, Hothorn, and Hornik 2008) and Rasch models in…”
A global assessment of the social and conservation outcomes of protected areas (2016, positive): “Variable importance thresholds were set as the value above the absolute value of the lowest negative-scoring variable (Strobl et al. 2009).”
Focused identification of germplasm strategy (FIGS) detects wheat stem rust resistance linked to environmental variables (2011, background): “Random forest (RF) RF is a type of recursive partitioning algorithm where the data is recursively split into groups of observations with similar response values, a procedure that does not require normality assumptions and deals well with a large number of variables (Strobl et al. 2009).”
Nitrogen fertilization reduces yield declines following no-till adoption (2015, methodology): “32 19 55 65 13 22 37 earch 183 (2015) 204–210 conditional random forest classification and regression approaches using the “cforest” procedure in the “party” package in R (Hothorn et al., 2006; Strobl et al., 2009).”
A Brief Review of Random Forests for Water Scientists and Practitioners and Their Recent History in Water Resources (2019, background): “3), with non-significant variables exhibiting randomly distributed VIMs around zero [71].”
Decision-tree analysis of factors influencing rainfall-related building structure and content damage (2014, methodology): “With such approaches, results are aggregated over an ensemble of trees, where each tree is based on random but realistic changes in the training data (Elith et al., 2008; Borisov, 2009; Strobl et al., 2009).”
Landslide susceptibility estimation by random forests technique: sensitivity and scaling issues (2013, background): “They are being increasingly used, however, also in environmental modelling (Prasad et al., 2006; Strobl et al., 2008; Bachmair and Weiler, 2012) and, even though only rarely, in landslide Nat.”