Article
Computer Science, Information Systems
Fa Zhu, Xingchi Chen, Xizhan Gao, Weidu Ye, Hai Zhao, Athanasios V. Vasilakos
Summary: This paper proposes a method called Constraint-weighted Support Vector Ordinal Regression (CWSVOR) to address the problem of constraint noises in ordinal regression. By introducing a constraint weight vector to control the influence of constraints on parallel hyperplanes, CWSVOR aims to mitigate the detrimental effects of constraint noises and shows superior performance on training sets corrupted by noises.
INFORMATION SCIENCES
(2023)
Article
Statistics & Probability
Shanghong Xie, Thaddeus Tarpey, Eva Petkova, R. Todd Ogden
Summary: This article proposes an approach to improve the accuracy of individualized treatment rules (ITRs) by using multiple kernel functions to describe the similarity of features. The method takes into account the heterogeneity of each data domain and combines data from multiple domains optimally. The approach can estimate optimal ITRs and identify the most important domains for determining ITRs.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2022)
Article
Computer Science, Information Systems
Zichen Zhang, Shifei Ding, Yuting Sun
Summary: This paper introduces a new method called multiple birth support vector regression (MBSVR), which constructs the regressor from multiple hyperplanes obtained by solving small quadratic programming problems, aiming for faster computation and better fitting precision.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Guoquan Li, Linxi Yang, Zhiyou Wu, Changzhi Wu
Summary: Proximal support vector machine (PSVM) is a variant of support vector machine (SVM) which aims to generate a pair of non-parallel hyperplanes for classification. Introducing l(0)-norm regularization in PSVM enables simultaneous selection of important features and removal of redundant features for classification. The proposed method utilizes a continuous nonconvex function and difference of convex functions algorithms (DCA) to solve the optimization problem efficiently.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Fa Zhu, Xingchi Chen, Shuo Chen, Wei Zheng, Weidu Ye
Summary: As a classical ordinal regression model, support vector ordinal regression (SVOR) finds parallel discriminant hyperplanes to maximize the minimal margins between different ranks. However, SVOR only considers minor patterns near the margin hyperplanes and ignores the contributions of other patterns. To address this issue, this paper proposes relative margin induced support vector ordinal regression (RMSVOR) models, which depict the margin between a pattern and a discriminant hyperplane based on relative margin information. Experimental results on various datasets show that RMSVOR outperforms previous ordinal regression models and canonical multi-class classification models.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Yafen Ye, Zhihu Xu, Jinhua Zhang, Weijie Chen, Yuanhai Shao
Summary: This paper proposes a twin support vector quantile regression (TSVQR) method to capture the heterogeneous and asymmetric information in modern data. TSVQR effectively depicts the heterogeneous distribution information with respect to all portions of data points using a quantile parameter. The method constructs two smaller sized quadratic programming problems to measure the distributional asymmetry between the lower and upper bounds at each quantile level. Experimental results show that TSVQR outperforms previous quantile regression methods in terms of capturing heterogeneous and asymmetric information effectively and efficiently in various datasets.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Chemistry, Analytical
Xihui Bian, Deyun Wu, Kui Zhang, Peng Liu, Huibing Shi, Xiaoyao Tan, Zhigang Wang
Summary: This study proposes a weighted multiscale support vector regression method based on variational mode decomposition for food and herb analysis. The method decomposes the spectra into discrete mode components, builds sub-models using support vector regression, and obtains the final prediction by averaging the predictions of the sub-models. Experimental results show that the method has potential in model accuracy.
Article
Computer Science, Artificial Intelligence
Xiaoming Wang, Shitong Wang, Zengxi Huang, Yajun Du
Summary: This paper introduces a novel method called sparse support vector machine guided by radius-margin bound (RMB-SSVM) to efficiently condense the basis vectors in support vector machines. By selecting basis vectors and learning corresponding coefficients with a criterion related to SVM's generalization ability, the RMB-SSVM model can yield better performance.
APPLIED SOFT COMPUTING
(2021)
Article
Computer Science, Artificial Intelligence
Huajuan Huang, Xiuxi Wei, Yongquan Zhou
Summary: This article reviews the recent developments in twin support vector regression (TSVR). It introduces the basic concepts and models of TSVR, summarizes the improved algorithms and applications in recent years, and analyzes the advantages and disadvantages of representative algorithms through experiments. The article also discusses the research conducted on TSVR.
Article
Computer Science, Artificial Intelligence
Quentin Klopfenstein, Samuel Vaiter
Summary: This paper investigates the addition of linear constraints to Support Vector Regression with a linear kernel, proving that the problem remains a semi-definite quadratic problem. A generalization of the Sequential Minimal Optimization algorithm is proposed to solve the optimization problem with linear constraints, showing convergence. Practical performance of this approach is demonstrated on simulated and real datasets, highlighting its usefulness compared to classical methods.
Article
Management
Zhiqiang Liao, Sheng Dai, Timo Kuosmanen
Summary: Nonparametric regression subject to convexity or concavity constraints is gaining popularity in various fields. The conventional convex regression method often suffers from overfitting and outliers. This paper proposes the convex support vector regression method to address these issues and demonstrates its advantages in prediction accuracy and robustness through numerical experiments.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2024)
Article
Computer Science, Artificial Intelligence
Deepak Gupta, N. Natarajan
Summary: The study introduces a novel density weighted approach for PLSTSVR to handle input samples in the presence of outliers, boosting the performance of PDWLSTSVR in terms of efficiency. The weights are determined with the help of k-Nearest Neighbour (k-NN) distance. Further, the proposed PDWLSTSVR outperforms the RF, ELM, LSSVR and PLSTSVR in terms of all evaluation measures when applied to the real-world application of predicting the UCS of rock samples.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Junyou Ye, Zhixia Yang, Mengping Ma, Yulan Wang, Xiaomei Yang
Summary: In this paper, a new regression method called epsilon-kernel-free soft quadratic surface support vector regression (epsilon-SQSSVR) is proposed. The method converts the regression problem into a classification problem and constructs an optimization problem based on maximizing the sum of relative geometrical margin of each training point. The model is nonlinear, kernel-free, and highly interpretable.
INFORMATION SCIENCES
(2022)
Article
Physics, Multidisciplinary
Huan Liu, Jiankai Tu, Chunguang Li
Summary: This paper proposes a distributed SVOR algorithm to solve ordinal regression problems in distributed environments. Theoretical analysis and experimental results demonstrate that the proposed method can achieve good performance in scenarios where privacy protection or centralized data processing is not feasible.
Article
Computer Science, Interdisciplinary Applications
Huan Luo, Stephanie German Paal
Summary: This paper proposes a novel and robust machine learning method, which formulates the problem as an optimization problem by combining locally weighted least-squares support vector machines for regression with a weight function. The proposed approach effectively reduces the interference of outliers and improves the predictive performance for various engineering problems.
ENGINEERING WITH COMPUTERS
(2023)
Article
Immunology
Maureen D. Goss, Jonathan L. Temte, Shari Barlow, Emily Temte, Cristalyne Bell, Jen Birstler, Guanhua Chen
Article
Cardiac & Cardiovascular Systems
Rachana D. Shah, Zheng-Zheng Tang, Guanhua Chen, Shi Huang, Jane F. Ferguson
NUTRITION METABOLISM AND CARDIOVASCULAR DISEASES
(2020)
Article
Statistics & Probability
Jared D. Huling, Maureen A. Smith, Guanhua Chen
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2020)
Article
Computer Science, Artificial Intelligence
David P. Azari, Brady L. Miller, Brian V. Le, Jacob A. Greenberg, Reginald C. Bruskewitz, Kristin L. Long, Guanhua Chen, Robert G. Radwin
Summary: This study used video-recorded hand motion data to create models predicting expert-rated performance on surgical motion scales, potentially enabling automated assessment and assistance in surgical training. The results showed that models predicting fluidity of motion and motion economy outperformed those for hand coordination and tissue handling.
IEEE TRANSACTIONS ON HUMAN-MACHINE SYSTEMS
(2021)
Article
Health Care Sciences & Services
Natalie Liu, Jen Birstler, Manasa Venkatesh, Lawrence Hanrahan, Guanhua Chen, Luke Funk
Summary: This study aimed to identify BMI cut points for predicting obesity-related comorbidities. Significant associations between increasing BMIs and the incidence of several comorbidities were found, with cut points identified for hyperlipidemia, coronary artery disease, hypertension, osteoarthritis, obstructive sleep apnea, and type 2 diabetes occurring when patients were overweight or barely met the criteria for class 1 obesity. Further research using national longitudinal data is needed to determine potential revisions to screening guidelines for appropriate comorbidities.
JOURNAL OF MEDICAL INTERNET RESEARCH
(2021)
Review
Oncology
Ronghui Xu, Guanhua Chen, Michael Connor, James Murphy
Summary: This article reviews different applications of incorporating individual patient variables into clinical research in oncology. The methods discussed range from traditional use of baseline covariates to generalize results, to considering treatment effects heterogeneity and individualized treatment rules. The article also discusses newer statistical research that is likely to impact future oncology research.
JOURNAL OF CLINICAL ONCOLOGY
(2022)
Article
Multidisciplinary Sciences
Cristalyne Bell, Maureen Goss, Jennifer Birstler, Emily Temte, Guanhua Chen, Peter Shult, Erik Reisdorf, Thomas Haupt, Shari Barlow, Jonathan Temte, Marianne Clemence, Emily Chenette
Summary: This study assessed the impact of clinical and laboratory factors on the sensitivity of fluorescence immunoassay. The results showed that factors such as the presence of an influenza-like illness, younger age, shorter time since onset, no co-detection, and the presence of nasal discharge were associated with higher sensitivity.
Article
Biology
Rui Chen, Guanhua Chen, Menggang Yu
Summary: This paper discusses how to estimate the average treatment effect (ATE) of a target population with both individual-level data from a source population and summary-level data from the target population. When there is heterogeneity in treatment effects, the ATE of the target population can differ from that of the source population due to covariate shift. Existing methods to adjust for covariate shift typically require individual covariates from a representative target sample. We propose a weighting approach using summary-level information from the target sample to adjust for possible covariate shift and achieve covariate balance in the source sample. Theoretical implications are supported by simulation studies and real-data application.
Article
Biology
Rui Chen, Jared D. Huling, Guanhua Chen, Menggang Yu
Summary: Learning individualized treatment rules is important in precision medicine, especially when the source population differs from the target population. This paper proposes a weighting framework to improve the generalizability of optimal individualized treatment rules by mitigating the impact of misspecification. The proposed method shows promising results in improving individualized treatment rule estimation for the target population.
Review
Oncology
Jens Eickhoff, Jen Zaborek, Guanhua Chen, Vikrant V. Sahasrabuddhe, Leslie G. Ford, Eva Szabo, KyungMann Kim
Summary: Early phase cancer prevention trials often fail to detect intervention effects. A systematic review of recently completed trials showed substantial differences between hypothesized and observed effect sizes, highlighting the need for careful planning of study design and sample size determination to detect meaningful intervention effects.
CANCER PREVENTION RESEARCH
(2023)
Article
Computer Science, Interdisciplinary Applications
Jifan Gao, Shilu He, Junjie Hu, Guanhua Chen
Summary: This paper presents a novel solution for predicting the relations between assessment and plan subsections in progress notes. The approach goes beyond standard transformer models by incorporating medical ontology and order information. The results show that this approach outperforms other systems in predicting the relationships between assessment and plan subsections.
JOURNAL OF BIOMEDICAL INFORMATICS
(2023)
Article
Statistics & Probability
Jared D. Huling, Noah Greifer, Guanhua Chen
Summary: Studying the causal effects of continuous treatments is crucial but limited to observational studies, leading to the issue of confounding. Weighting approaches are employed to address confounding, but they are sensitive to model misspecification when it comes to continuous treatments. In this article, the authors propose a measure to eliminate confounding and a new model-free method for weight estimation. The theoretical properties and empirical effectiveness of the proposed approach are examined, demonstrating its robustness in various scenarios.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2023)
Article
Medicine, Research & Experimental
Timothy Bergquist, Marie Wax, Tellen D. Bennett, Richard A. Moffitt, Jifan Gao, Guanhua Chen, Amalio Telenti, M. Cyrus Maher, Istvan Bartha, Lorne Walker, Benjamin E. Orwoll, Meenakshi Mishra, Joy Alamgir, Bruce L. Cragin, Christopher H. Ferguson, Hui-Hsing Wong, Anne Deslattes Mays, Leonie Misquitta, Kerry A. DeMarco, Kimberly L. Sciarretta, Sandeep A. Patel
Summary: COVID-19 continues to pose a burden on the pediatric population due to various factors. There is a need to proactively identify pediatric patients at risk of severe COVID-19 and other respiratory infectious diseases. However, there is currently no nationwide capability for developing validated computational tools to identify these patients using real-world data.
JOURNAL OF CLINICAL AND TRANSLATIONAL SCIENCE
(2023)
Article
Biotechnology & Applied Microbiology
Qilin Hong, Guanhua Chen, Zheng-Zheng Tang
Summary: PhyloMed is a phylogeny-based mediation analysis method that addresses the challenges of compositional and high-dimensional microbiome data. It discovers mediation signals by analyzing subcompositions defined on the phylogenetic tree, producing well-calibrated mediation test p-values and higher discovery power than existing methods.
Article
Mathematical & Computational Biology
Rui Chen, Guanhua Chen, Menggang Yu
Summary: Generalizing population level causal quantities from a source population to a target population can be difficult and unreliable when there are heterogenous causal effects and differences in subject characteristics. To address this issue, we propose a generalizability score that can be used as a yardstick to select target subpopulations and prevent biases associated with inadvertent access to outcome information.
