Feature Importance Measurement based on Decision Tree Sampling

Chao Huang, Diptesh Das, Koji Tsuda
7/25/2023
cs.LG

Official code: tsudalab/DT-sampler (5 stars, last commit 2023-09-02).

Abstract

Random forest is effective for prediction tasks but the randomness of tree generation hinders interpretability in feature importance analysis. To address this, we proposed DT-Sampler, a SAT-based method for measuring feature importance in tree-based model. Our method has fewer parameters than random forest and provides higher interpretability and stability for the analysis in real-world problems. An implementation of DT-Sampler is available at https://github.com/tsudalab/DT-sampler.

View on arXivView PDF

Code Implementations(1)

tsudalab/DT-samplerOfficial
51PythonJul 17, 20233 years ago

Similar Papers

Accurate estimation of feature importance faithfulness for tree models

81% match2024

Local MDI+: Local Feature Importances for Tree-Based Models

79% match2025

Conditional Feature Importance with Generative Modeling Using Adversarial Random Forests

78% match1 repo42025

CID: Measuring Feature Importance Through Counterfactual Distributions

78% match2025

The Berkelmans-Pries Feature Importance Method: A Generic Measure of Informativeness of Features

77% match2023

Enhancing Variable Importance in Random Forests: A Novel Application of Global Sensitivity Analysis

77% match2024

Cite this paper

@article{huang2023feature,
  title  = {Feature Importance Measurement based on Decision Tree Sampling},
  author = {Chao Huang and Diptesh Das and Koji Tsuda},
  year   = {2023},
  eprint = {2307.13333},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2307.13333v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2307.13333.svg)](https://codeofpaper.com/paper/2307.13333)

Discussion