☆ 4.1 Article

Conditional feature importance for mixed data

ASTA-ADVANCES IN STATISTICAL ANALYSIS (2023)

相关参考文献

注意：仅列出部分参考文献，下载原文获取全部文献信息。

Article Computer Science, Artificial Intelligence

Model-agnostic feature importance and effects with dependent features: a conditional subgroup approach

Christoph Molnar et al.

Summary: The interpretation of feature importance in machine learning models becomes challenging when features are dependent. Permutation feature importance (PFI) can be misleading due to its ignorance of such dependencies. To address this issue, this paper proposes a new sampling mechanism based on permutations in conditional subgroups, which provides a transparent and comprehensible conditional assessment of feature importance. This approach not only offers a simple and effective estimator of conditional PFI, but also enables fine-grained interpretation within the conditional subgroups.

DATA MINING AND KNOWLEDGE DISCOVERY (2023)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Grouped feature importance and combined features effect plot

Quay Au et al.

Summary: This paper focuses on the assessment of feature group importance and the visualization of feature group effects in interpretable machine learning algorithms, and proposes a sequential procedure based on importance and a combined features effect plot as solutions.

DATA MINING AND KNOWLEDGE DISCOVERY (2022)

添加到收藏夹

Article Statistics & Probability

Metropolized Knockoff Sampling

Stephen Bates et al.

Summary: This article introduces the concept of Model-X knockoffs and techniques for knockoff generation, as well as how to use knockoffs to control false positives, speed up computations, and apply to challenging models.

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION (2021)

添加到收藏夹

Article Biology

Nonparametric variable importance assessment using machine learning techniques

Brian D. Williamson et al.

Summary: In this work, a variable importance measure that can be used with any regression technique, regardless of the technique used, is studied. The study discusses how to flexibly estimate the importance of a single feature or group of features using machine learning techniques. Through simulations and a case study, it is shown that the proposal has good practical operating characteristics and effects.

BIOMETRICS (2021)

添加到收藏夹

Review Physics, Multidisciplinary

Explainable AI: A Review of Machine Learning Interpretability Methods

Pantelis Linardatos et al.

Summary: Recent advances in artificial intelligence have led to widespread industrial adoption, with machine learning systems demonstrating superhuman performance. However, the complexity of these systems has made them difficult to explain, hindering their application in sensitive domains. Therefore, there is a renewed interest in the field of explainable artificial intelligence.

ENTROPY (2021)

添加到收藏夹

Article Computer Science, Theory & Methods

Unrestricted permutation forces extrapolation: variable importance requires at least one more model, or there is no free variable importance

Giles Hooker et al.

Summary: This paper discusses the drawbacks of using permute-and-predict methods for interpreting black box functions and introduces arguments against their use. The authors highlight the possibility of misleading diagnostics when dealing with strong feature dependencies and advocate for alternative measures involving additional modeling.

STATISTICS AND COMPUTING (2021)

添加到收藏夹

Article Mathematical & Computational Biology

Sequential knockoffs for continuous and categorical predictors: With application to a large psoriatic arthritis clinical trial pool

Matthias Kormaksson et al.

Summary: This paper introduces a Knockoffs framework for controlling the false discovery rate in variable selection and proposes a new algorithm and method, validated through extensive simulations. The findings have significant implications for medical practice and other fields where variable selection is crucial.

STATISTICS IN MEDICINE (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Testing conditional independence in supervised learning algorithms

David S. Watson et al.

Summary: The proposed method introduces a new estimator called Conditional Predictive Impact (CPI) for measuring the association between features and outcomes under reduced feature sets. Through various algorithm tests and simulations, it has been demonstrated that CPI performs favorably compared to alternative methods.

MACHINE LEARNING (2021)

添加到收藏夹

Article Statistics & Probability

Deep Knockoffs

Yaniv Romano et al.

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION (2020)

添加到收藏夹

Article Statistics & Probability

THE HARDNESS OF CONDITIONAL INDEPENDENCE TESTING AND THE GENERALISED COVARIANCE MEASURE

Rajen D. Shah et al.

ANNALS OF STATISTICS (2020)

添加到收藏夹

Article Statistics & Probability

Visualizing the effects of predictor variables in black box supervised learning models

Daniel W. Apley et al.

JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY (2020)

添加到收藏夹

Article Computer Science, Artificial Intelligence

From local explanations to global understanding with explainable AI for trees

Scott M. Lundberg et al.

NATURE MACHINE INTELLIGENCE (2020)

添加到收藏夹

Article Biology

Gene hunting with hidden Markov model knockoffs

M. Sesia et al.

BIOMETRIKA (2019)

添加到收藏夹

Review Genetics & Heredity

Review of Causal Discovery Methods Based on Graphical Models

Clark Glymour et al.

FRONTIERS IN GENETICS (2019)

添加到收藏夹

Article Statistics & Probability

Distribution-Free Predictive Inference for Regression

Jing Lei et al.

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION (2018)

添加到收藏夹

Article Statistics & Probability

Panning for gold: model-X' knockoffs for high dimensional controlled variable selection

Emmanuel Candes et al.

JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY (2018)

添加到收藏夹

Article Computer Science, Interdisciplinary Applications

Feature Selection with the Boruta Package

Miron B. Kursa et al.

JOURNAL OF STATISTICAL SOFTWARE (2010)

添加到收藏夹

Article Statistics & Probability

Regularization and variable selection via the elastic net

H Zou et al.

JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY (2005)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Random forests

L Breiman

MACHINE LEARNING (2001)

添加到收藏夹

Article Statistics & Probability

Greedy function approximation: A gradient boosting machine

JH Friedman

ANNALS OF STATISTICS (2001)

添加到收藏夹

© Peeref 2019-2024. All rights reserved.