Article
Statistics & Probability
Leo Miolane, Andrea Montanari
Summary: The Lasso is a popular regression method for high-dimensional problems, with statistical properties related to soft-thresholding denoisers. The method can be used to evaluate the performance of various data-driven procedures and has been shown to be effective in dealing with Gaussian noise.
ANNALS OF STATISTICS
(2021)
Article
Statistics & Probability
Maryclare Griffin, Peter D. Hoff
Summary: In many regression settings, the unknown coefficients may have known structure and be sparse. However, commonly used priors and penalties do not encourage both structured and sparse estimates. This article develops structured shrinkage priors that allow for correlated coefficients while maintaining sparsity, and presents a computational approach to overcome the challenges associated with these priors. The results demonstrate the effectiveness of these priors in introducing structure while preserving sparsity.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2023)
Article
Biochemical Research Methods
Ayyuce Begum Bektas, Cigdem Ak, Mehmet Gonen
Summary: With the increasing sizes of computational biology datasets, previous kernel-based machine learning algorithms have failed to provide satisfactory interpretability. To address this issue, we propose a fast and efficient multiple kernel learning algorithm that can extract significant information from genomic data. Our experiments demonstrate that the algorithm outperforms baseline methods while using only a small fraction of input features, and it has the potential to discover new biomarkers and therapeutic guidelines.
Article
Biochemical Research Methods
Akio Onogi, Aisaku Arakawa
Summary: An R package has been developed to implement multiple linear learners in a single model. It uses fast algorithms to obtain solutions and is useful for incorporating multimodal and high-dimensional explanatory variables in regression models.
Article
Mathematics, Applied
Qianru Liu, Rui Wang, Yuesheng Xu, Mingsong Yan
Summary: This paper studies a regularization problem that combines a convex fidelity term and a regularization term based on the 21 norm and linear transform. The empirical results demonstrate that the regularization with the 21 norm promotes sparsity in the solution. The main goal of this paper is to theoretically understand the impact of the regularization parameter on the sparsity of the solution. The paper establishes a characterization of the sparsity under the transform matrix of the solution and proposes iterative algorithms for determining the regularization parameter and its corresponding solution with a prescribed sparsity level.
Article
Multidisciplinary Sciences
Eric J. Ward, Kristin Marshall, Mark D. Scheuerell
Summary: This study investigates the potential benefits of using regularized priors for estimating inter-specific interactions, and demonstrates their effectiveness through simulations and a case study.
Article
Mathematical & Computational Biology
Nadim Ballout, Cedric Garcia, Vivian Viallon
Summary: This study compared two regression analysis methods for disease subtypes in case-control studies, one based on data shared lasso and the other on L-1-norm penalized multi-nomial logistic regression. Experimental results showed that the non-symmetric formulation method is not recommended when homogeneity is high.
Article
Multidisciplinary Sciences
Pathum Kossinna, Weijia Cai, Xuewen Lu, Carrie S. Shemanko, Qingrun Zhang
Summary: This article introduces a tool called SCOPE, which integrates bootstrapped least absolute shrinkage and selection operator and coexpression analysis to obtain stable results that are insensitive to variations in the data. By applying SCOPE to cancer expression datasets, core genes capturing interaction effects in crucial pan-cancer pathways related to genome instability and DNA damage response were identified, highlighting the pivotal role of CD63 as an oncogenic driver and potential therapeutic target in kidney cancer.
Article
Chemistry, Multidisciplinary
Xin Qiao, Yoshikazu Kobayashi, Kenichi Oda, Katsuya Nakamura
Summary: A novel acoustic emission (AE) tomography algorithm based on Lasso regression (LASSO) is developed for non-destructive testing (NDT). The algorithm eliminates the deficiencies of the conventional AE tomography method and reconstructs equivalent velocity distribution with fewer event data. The study demonstrates the applicability of the LASSO algorithm in AE tomography and the elimination of shadow parts in resultant elastic velocity distributions.
APPLIED SCIENCES-BASEL
(2022)
Article
Mathematics
Antonis Christou, Andreas Artemiou
Summary: In this study, we propose a sparse version of Support Vector Regression (SVR) that achieves sparsity in function estimation through regularization. We introduce an adaptive L-0 penalty with a ridge structure, which does not increase computational complexity. Additionally, we adopt an alternative approach based on a similar proposal in the Support Vector Machine (SVM) literature. Numerical studies confirm the effectiveness of our novel approach, which focuses on variable selection rather than support vector selection.
Article
Engineering, Multidisciplinary
Xiannan Han, Guobin Chang, Nanshan Zheng, Shubi Zhang
Summary: In this study, the traditional Hardy function interpolation method was improved by using L1-norm regularization, which automatically selects the best model and achieves sparse performance. The results showed that in the East direction, the interpolation effect of the L1-norm regularization method was significantly better than that of the Tikhonov regularization method, while there was a slight decrease in the North and Up directions.
MEASUREMENT SCIENCE AND TECHNOLOGY
(2021)
Article
Geochemistry & Geophysics
Xiangfei Shen, Haijun Liu, Xinzheng Zhang, Kai Qin, Xichuan Zhou
Summary: The study introduces a local-global sparse regression unmixing method, which combines local sparsity regularization and global sparsity regularization to effectively estimate the abundance of a given image. Experimental results demonstrate the effectiveness of the proposed algorithm.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Statistics & Probability
Cheng Zhang, Vu Dinh, Frederick A. Matsen
Summary: This research introduces adaptive-LASSO-type regularization estimators to identify zero-length branches in phylogenetic trees, proving regularization to be a practical method in phylogenetics. This approach helps uncover special features in densely sampled phylogenetic trees, such as sampled ancestors and polytomies.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Article
Multidisciplinary Sciences
Ritwik Vashistha, Zubdahe Noor, Shibasish Dasgupta, Jie Pu, Shibing Deng
Summary: This study evaluated different variable selection methods for discovering prognostic and predictive biomarkers in patients with advanced urothelial cancer who received first-line platinum-based chemotherapy. The results demonstrated the limitations of existing methods in the presence of high collinearity. Potentially significant biomarkers were identified in the JAVELIN Bladder 100 dataset.
SCIENTIFIC REPORTS
(2023)
Article
Biochemical Research Methods
Michael Komodromos, Eric O. Aboagye, Marina Evangelou, Sarah Filippi, Kolyan Ray
Summary: This paper proposes an interpretable and scalable Bayesian proportional hazards model, referred to as sparse variational Bayes, for analyzing high-dimensional sparse survival data. The proposed method overcomes the high computational cost of traditional methods and offers a mechanism for variable selection via posterior inclusion probabilities. Extensive simulations demonstrate the comparable or better performance of the proposed method compared to state-of-the-art Bayesian variable selection methods.
Article
Behavioral Sciences
Heather A. Eicher-Miller, Saul Gelfand, Youngha Hwang, Edward Delp, Anindya Bhadra, Jiaqi Guo
Article
Statistics & Probability
Anindya Bhadra, Jyotishka Datta, Yunfan Li, Nicholas G. Polson
INTERNATIONAL STATISTICAL REVIEW
(2020)
Article
Nutrition & Dietetics
Alexandra E. Cowan, Shinyoung Jun, Janet A. Tooze, Heather A. Eicher-Miller, Kevin W. Dodd, Jaime J. Gahche, Patricia M. Guenther, Johanna T. Dwyer, Nancy Potischman, Anindya Bhadra, Regan L. Bailey
Review
Nutrition & Dietetics
Marah Aqeel, Anna Forster, Elizabeth A. Richards, Erin Hennessy, Bethany McGowan, Anindya Bhadra, Jiaqi Guo, Saul Gelfand, Edward Delp, Heather A. Eicher-Miller
Article
Statistics & Probability
Yunfan Li, Jyotishka Datta, Bruce A. Craig, Anindya Bhadra
Summary: Seemingly unrelated regression is a flexible framework for regressing multiple correlated responses on multiple predictors. The use of horseshoe priors on both the mean vector and the inverse covariance matrix addresses the challenges of inferring both parameters simultaneously in a Bayesian framework.
JOURNAL OF MULTIVARIATE ANALYSIS
(2021)
Article
Mathematical & Computational Biology
Ritwik Bhaduri, Ritoban Kundu, Soumik Purkayastha, Michael Kleinsasser, Lauren J. Beesley, Bhramar Mukherjee, Jyotishka Datta
Summary: The study suggests that considering false negative rates of diagnostic tests for severe acute respiratory coronavirus 2 and selection bias due to prioritized testing, and extending the widely used SEIR model can improve the accuracy of COVID-19 transmission dynamics modeling. Analyzing data from the first two waves of the pandemic in India, the study provides estimates of undetected infections and deaths, and demonstrates the impact of misclassification and selection on future infection prediction and R0 estimation.
STATISTICS IN MEDICINE
(2022)
Article
Computer Science, Information Systems
Alfieri Ek, Grant Drawve, Samantha Robinson, Jyotishka Datta
Summary: Law enforcement agencies are increasingly using spatial analysis to identify patterns of outcomes. However, there has been little progress in the statistical modeling of mental health events in Little Rock, Arkansas. In this article, insights into the spatial nature of mental health data from 2015 to 2018 in Little Rock, Arkansas are provided. Different models are used and their relative predictive performances are presented. The findings have the potential to assist law enforcement agencies and the city in resource allocation.
ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION
(2023)
Article
Biodiversity Conservation
John G. Jelesko, Kyla Thompson, Noah Magerkorth, Elizabeth Verteramo, Hannah Becker, Joy G. Flowers, Jonathan Sachs, Jyotishka Datta, Jordan Metzgar
Summary: Urushiol, produced by poison ivy, causes millions of cases of delayed contact allergenic dermatitis in North America annually. Avoiding poison ivy plant material is recommended to prevent symptoms. However, the variable leaf shape of poison ivy makes accurate identification difficult.
PLANTS PEOPLE PLANET
(2023)
Article
Engineering, Electrical & Electronic
Ksheera Sagar, Anindya Bhadra
Summary: The horseshoe prior, a state of the art approach to Bayesian sparse signal recovery, is defined as a half Cauchy scale mixture of normal. In this study, we provide a new representation of the horseshoe density as a scale mixture of the Laplace density, explicitly identifying the mixing measure. The complete monotonicity of the horseshoe density and the strong concavity of the corresponding penalty are established, using the Bernstein-Widder theorem and a result due to Bochner. Moreover, the equivalence between local linear approximation and expectation-maximization algorithms for finding the posterior mode under the horseshoe penalized regression is proven, and the resultant estimate is shown to be sparse.
IEEE SIGNAL PROCESSING LETTERS
(2022)
Article
Criminology & Penology
Grant Drawve, Casey T. Harris, Shaun A. Thomas, Jyotishka Datta, Jack Cothren
Summary: This study focuses on criminal incidents reported to the National Incident Based Reporting System in Arkansas in 2016, aiming to showcase the benefits of collecting address-identified NIBRS data for Arkansas and other states. By comparing statewide NIBRS data with address-level data for a specific city, the study illustrates spatial variation in crime occurrence at different levels of analysis.
CRIME & DELINQUENCY
(2021)
Article
Statistics & Probability
Anindya Bhadra, Jyotishka Datta, Nicholas G. Polson, Brandon T. Willard
Summary: This paper introduces an alternative method for feature subset selection called the horseshoe regularization penalty, which shows superior theoretical and computational performance compared to existing methods. The distinguishing feature is the probabilistic representation of the penalty, enabling efficient optimization algorithms and uncertainty quantification.
SANKHYA-SERIES B-APPLIED AND INTERDISCIPLINARY STATISTICS
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Souradip Chakraborty, Ekansh Verma, Saswata Sahoo, Jyotishka Datta
20TH IEEE INTERNATIONAL CONFERENCE ON DATA MINING WORKSHOPS (ICDMW 2020)
(2020)
Article
Statistics & Probability
Anindya Bhadra, Jyotishka Datta, Nicholas G. Poison, Brandon T. Willard
SANKHYA-SERIES A-MATHEMATICAL STATISTICS AND PROBABILITY
(2020)
