Article
Biology
S. Klaassen, J. Kueck, M. Spindler, V Chernozhukov
Summary: This paper investigates the uniform inference on high-dimensional graphical models under approximate sparsity, and demonstrates how to estimate and recover the graphical models using modern machine learning methods. The paper establishes uniform estimation rates and sparsity guarantees for the square-root lasso estimator in a random design, and demonstrates its good performance through comprehensive simulations.
Article
Statistics & Probability
Eftychia Solea, Holger Dette
Summary: This article focuses on constructing nonparametric undirected graphical models for high-dimensional functional data. A more flexible model is proposed, replacing the linearity assumption with an arbitrary additive form. The use of functional principal components and a group lasso penalty allows for estimation of the relevant edges of the graph. Statistical guarantees are established, and empirical performance is evaluated through simulation studies and a real data application.
ELECTRONIC JOURNAL OF STATISTICS
(2022)
Article
Automation & Control Systems
Snigdha Panigrahi, Peter W. MacDonald, Daniel Kessler
Summary: In this paper, a consistent, post-selective Bayesian method is proposed to address the issues in Group LASSO selection. The method derives a likelihood adjustment factor and an approximation to eliminate bias from the selection of groups.
JOURNAL OF MACHINE LEARNING RESEARCH
(2023)
Article
Automation & Control Systems
Fei Wang, Ling Zhou, Lu Tang, Peter X. K. Song
Summary: The paper introduces a method of contraction and expansion (MOCE) for simultaneous inference after model selection in high-dimensional linear regression models, with theoretical guarantees and stable coverage.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Mathematics
Francisco Javier Diez, Manuel Arias, Jorge Perez-Martin, Manuel Luque
Summary: OpenMarkov is an open-source software tool designed for probabilistic graphical models, primarily in medicine but also used in other fields and education in over 30 countries. This paper explains how OpenMarkov can be used as a pedagogical tool to teach the main concepts of Bayesian networks and influence diagrams, as well as various inference algorithms.
Article
Statistics & Probability
Guillaume Marrelec, Alain Giron, Laura Messio
Summary: The study investigates a specific Gaussian graphical model and finds that pairwise correlation decays exponentially with distance, while also analyzing the difference between finite and infinite cases as the number of variables tends to infinity.
STATISTICS & PROBABILITY LETTERS
(2021)
Article
Automation & Control Systems
Hao Dong, Yuedong Wang
Summary: This paper presents a nonparametric neighborhood selection method under a unified framework for mixed data, which performs well in simulations and real data examples.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Mathematics
Claudia Angelini, Daniela De Canditiis, Anna Plaksienko
Summary: In this paper, we propose a joint data estimation method called jewel, using group Lasso penalty and establishing the consistency property of the estimator.
Article
Engineering, Mechanical
Daniel Correia, Daniel N. Wilke, Stephan Schmidt
Summary: Causal models are important in engineering, and both physics-based models and data-driven models are widely used. However, capturing complex forms and maintaining interpretability are challenges for causal models. This work proposes a method for automating the selection of functional forms of causal relationships in the context of counterfactual inference, using a sensible library of basis functions and enforcing sparsity for interpretability.
PROBABILISTIC ENGINEERING MECHANICS
(2022)
Article
Mathematics
Claudia Angelini, Daniela De Canditiis, Anna Plaksienko
Summary: This paper addresses the problem of estimating graphical models of conditional dependencies between variables from multiple datasets under Gaussian settings. The proposed jewel 2.0 method improves upon the previous version by modeling commonality and class-specific differences in the graph structures and incorporating a stability selection procedure to reduce false positives. The performance of jewel 2.0 is demonstrated through simulated and real data examples, and the method is implemented in the R package jewel.
Article
Mathematics
Lingju Chen, Shaoxin Hong, Bo Tang
Summary: This study focuses on the identification and estimation of graphical models with nonignorable nonresponse. The proposed method introduces a new observable variable to identify the mean of response for the unidentifiable model and suggests a simulation imputation approach to estimate the marginal mean of response. The root N-consistent mean estimators show effectiveness in finite sample simulations and a real data example is used to illustrate the methodology.
JOURNAL OF MATHEMATICS
(2021)
Article
Automation & Control Systems
Nguyen Thi Kim Hue, Monica Chiogna
Summary: Motivated by the complex interactions between genes in modeling biological processes, a new algorithm called PC-LPGM is proposed to learn the structure of undirected graphical models over discrete variables. The theoretical consistency of PC-LPGM and its robustness to model misspecification are proven, with extensive simulation studies and biological validation confirming its performance in recovering true graph structures.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Mechanics
Christoph Feinauer, Carlo Lucibello
Summary: Pairwise models like the Ising model or the generalized Potts model have been successfully applied in various fields, while the problem of inverse statistical mechanics aims to infer the parameters of such models from observed data. One open question is how to train these models when data contain higher-order interactions not present in the pairwise model. Proposed a hybrid model approach combining pairwise models and neural networks, showing significant improvements in reconstructing pairwise interactions. Results indicate that hybrids models can retain advantages of both simple interpretable models and complex black-box models.
JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT
(2021)
Article
Mechanics
A. Fanthomme, F. Rizzato, S. Cocco, R. Monasson
Summary: Understanding the role of regularization in statistical inference is important, and well-chosen regularization schemes can significantly improve the quality of inferred models. This study focuses on the L-2 regularization in maximum a posteriori (MAP) inference of generative pairwise graphical models. Analytical calculations and numerical experiments are used to examine the likelihoods of different data sets under various regularization strengths. The results show that the test and generated likelihoods have close values at their maximum, and the optimal regularization strength is related to the magnitude of couplings in the underlying network.
JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT
(2022)
Article
Mathematics, Interdisciplinary Applications
J. Jongerling, S. Epskamp, D. R. Williams
Summary: In the network approach to psychopathology, psychological constructs are conceptualized as networks of interacting components. This study compares estimation methods for symptom networks and finds that the Bayesian GLASSO performed better than the frequentist GLASSO in several measures of bias and specificity.
MULTIVARIATE BEHAVIORAL RESEARCH
(2023)