Interpreting tree ensembles with inTrees

By Houtao Deng. 2014

272 citations recorded by the source corpus.

Tree ensembles such as random forests and boosted trees are accurate but difficult to understand. In this work, we provide the interpretable trees (inTrees) framework that extracts, measures, prunes, selects, and summarizes rules from a tree ensemble, and calculates frequent variable interactions. The inTrees framework can be applied to multiple types of tree ensembles, e.g., random forests, regularized random forests, and boosted trees. We implemented the inTrees algorithms in the “inTrees” R package.

Subfields: Artificial Intelligence, Economics and Econometrics, Safety, Robustness & Interpretability

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 Economics and Econometrics

In Safety, Robustness & Interpretability

Citations in the corpus

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “Another approach for simplification is that in [118], in which authors create a Simplified Tree Ensemble Learner (STEL).”
A Survey of Methods for Explaining Black Box Models (2018, methodology): “In Reference [25] the complexity of a rule (and thus of an explanation) is measured by the length of the rule condition, defined as the number of attribute-value pairs in the condition.”
Interpretability via Model Extraction (2017, methodology): “Previous model extraction approaches have focused on specic model families [10, 28, 29], enabling them to leverage the internal structure of the model.”
Making Tree Ensembles Interpretable (2016, background): “The most relevant study compared with ours would be inTrees (interpretable trees) proposed by Deng (2014).”
Interpreting Blackbox Models via Model Extraction (2017, methodology): “Previous model extraction approaches have focused on interpreting speciﬁc families of models such as random forests [45, 18, 46], enabling them to leverage domain-speciﬁc knowledge about the internal structure of the model.”
Explainable Artificial Intelligence: a Systematic Review (2020, background): “inTrees [296] extracts, prunes and selects rules from a tree ensemble.”
Counterfactual Explanations for Machine Learning: A Review (2020, background): “Various papers have developed techniques to explain neural networks and tree ensembles using single decision tree [25, 34, 70] and rule sets [13, 28].”
GENESIM: genetic extraction of a single, interpretable model (2016, methodology): “Deng [4] introduced STEL, which converts an ensemble into an ordered rule list using the following steps.”
Making Tree Ensembles Interpretable: A Bayesian Model Selection Approach (2016, methodology): “The ﬁrst three are the tree ensemble simpliﬁcation methods: BATrees (Breiman and Shang, 1996), inTrees (Deng, 2014), and Node Harvest (NH) (Meinshausen, 2010).”
Ensemble learning: A survey (2018, background): “A simplified tree ensemble learner (STEL; Deng, 2014) provides the ability to extract rules from an ensemble of trees and search for frequent variable interactions.”
iForest: Interpreting Random Forests via Visual Analytics (2019, background): “Surrogate models are often simple and interpretable, such as decision trees [48, 60], decision rules [16, 23], and decision sets [32].”
Principles and Practice of Explainable Machine Learning (2020, background): “• Another approach to simplify RFs is discussed in [24].”
Discovery of cell-type specific DNA motif grammar in cis-regulatory elements using random Forest (2018, methodology): “In particular, we use a rule mining approach that is implemented in an R package called inTrees [31], which can extract, measure, prune and select rules from a RF.”
On sparse ensemble methods: An application to short-term predictions of the evolution of COVID-19 (2021, unknown): “The first two properties relate to the transparency nd interpretability of the ensemble, Deng (2019) and Florez-Lopez nd Ramon-Jeronimo (2015) , while the third one relates to the per- ormance of the ensemble in critical groups.”
Counterfactual Explanations and Algorithmic Recourses for Machine Learning: A Review (2020, unknown): “Various papers have developed techniques to explain neural networks and tree ensembles using single decision tree [65, 83, 184] and rule sets [14, 76].”
Techniques for interpretable machine learning (2018, methodology): “rediction performance, and the behavior of which is much easier to understand. For instance, the ensemble of decision trees can be transformed into a single decision tree [36] or an ordered rule list [9]. Moreover, a DNN is utilized to train a decision tree which mimics the input-output function cap”