Random survival forests
By Hemant Ishwaran, Udaya B. Kogalur, Eugene H. Blackstone, Michael S. Lauer. 2008
2,820 citations recorded by the source corpus.
We introduce random survival forests, a random forests method for the analysis of right-censored survival data. New survival splitting rules for growing survival trees are introduced, as is a new missing data algorithm for imputing missing data. A conservation-of-events principle for survival forests is introduced and used to define ensemble mortality, a simple interpretable measure of mortality that can be used as a predicted outcome. Several illustrative examples are given, including a case study of the prognostic implications of body mass for individuals with coronary artery disease. Computations for all examples were implemented using the freely available R-software package, randomSurvivalForest.
Subfields: Demography, Statistics and Probability
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Demography
- Hemant Ishwaran
- Ulrike Grömping
- S. van Buuren
- Chandan K. Reddy
- Yan Li
- Ping Wang
- Mihaela van der Schaar
- Ahmed M. Alaa
In Statistics and Probability
- Mihaela van der Schaar
- Ahmed M. Alaa
- Gérard Biau
- Erwan Scornet
- Ulrike Grömping
- Hemant Ishwaran
- S. van Buuren
- Hatem Ben-Ameur
Citations in the corpus
- Uniform Convergence of Random Forests via Adaptive Concentration (2015, neutral): “Notable extensions of the original random forest algorithm include online random forests [15, 34], random survival forests [23] and Bayesian alternatives to random forests [13].”
- DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network (2016, background): “Another popular machine learning approach to modeling patients’ hazard function is the random survival forest (RSF) [15, 16].”
- Data mining in clinical big data: the frequently used databases, steps, and methodological models (2021, background): “Additionally, random survival forests have recently been developed to analyse survival data, especially right-censored survival data [49, 50], which can help researchers conduct survival analyses in clinical oncology and help develop personalized treatment regimens that benefit patients [51].”
- Variable importance in binary regression trees and forests (2007, methodology): “In random forests, for example, VIMP for a variable x v is the difference between prediction error when x v is noised up by permuting its value randomly, compared to prediction error under the original predictor [2, 4, 5, 6].”
- A review of survival trees (2011, methodology): “The second aggregation method is the random survival forest (RSF) approach of Ishwaran et al. (2008) as implemented in the package randomSurvivalForest.”
- Random Forests, Decision Trees, and Categorical Predictors: The "Absent Levels" Problem (2017, background): “As pointed out in Ishwaran et al. (2008): Although surrogate splitting works well for trees, the method may not be well suited for forests.”
- scikit-survival: A Library for Time-to-Event Analysis Built on Top of scikit-learn (2020, background): “…many successful ideas from machine learning have been adapted for time-to-event analysis, such as gradient boosting (Ridgeway, 1999; Hothorn et al., 2006), random forests (Ishwaran et al., 2008), and support vector machines (Van Belle et al., 2007; Evers and Messow, 2008; Pölsterl et al., 2015).”
- Language Models are Few-shot Learners for Prognostic Prediction (2023, background): “…stored in tabular form, current mainstream machine learning approaches (Topol, 2019; Rajkomar et al., 2019) to cancer prognosis (Chowell et al., 2021) are still tree-based ensemble models such as boosting (Chen and Guestrin, 2016; Ke et al., 2017) and bagging (Breiman, 2004; Ishwaran et al., 2019).”
- Machine Learning for Survival Analysis (2019, methodology): “The Random Survival Forest (RSF) (Ishwaran et al. 2008) extended Breiman’s random forest method by using a forest of survival trees for prediction.”
- DeepHit: A Deep Learning Approach to Survival Analysis With Competing Risks (2018, background): “Recently developed survival models include random survival forests (Ishwaran et al. 2008), deep exponential families (R.”
- The revival of the Gini importance? (2018, methodology): “2 Power study To evaluate the combination of the AIR importance with the testing procedure by Janitza et al. (2016), we analyzed the type I error and power on two gene expression and one GWAS dataset.”
- Case-Specific Random Forests (2016, methodology): “The popularity of random forests (RFs) is reflected by its extension and incorporation in other methodology, such as multivariate random forests (De’ath 2010; Segal and Xiao 2011), quantile regression forests (Meinshausen 2006), enriched random forests for microarray analysis (Amaratunga, Cabrera, a”
- Improving Post-Click User Engagement on Native Ads via Survival Analysis (2016, methodology): “Survival Random Forest [15] is an ensemble tree method for the analysis of right-censored 4 lifetime data.”
- An interferon-related gene signature for DNA damage resistance is a predictive marker for chemotherapy and radiation for breast cancer (2008, methodology): “RSF is a non-parametric ensemble partitioning tree method for survival data that automatically estimates non-linear effects for variables and multi-way interactions between variables, and is capable of imputing missing data (14).”