Article
Computer Science, Artificial Intelligence
Hanyu E, Ye Cui, Witold Pedrycz, Zhiwu Li
Summary: This article aims to develop a framework for aggregating rule-based models by constructing a collection of models through random sampling and optimizing the design and aggregation using machine learning mechanisms. Experimental results demonstrate that the performance of the aggregated models outperforms the monolithically developed models.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2023)
Article
Economics
Toni Duras, Farrukh Javed, Kristofer Mansson, Paer Sjolander, Magnus Soderberg
Summary: Regulatory agencies often use nonparametric data envelopment analysis (DEA) to assess the efficiency of electricity providers. Traditional variable selection techniques are challenged by high-dimensional data in the era of big data. This study introduces new machine learning methods for energy market regulators, proposing a two-step approach that utilizes adaptive least absolute shrinkage and selection operator (ALASSO) in the first step and the selected variables in a DEA model in the second step. The results show that different machine learning techniques perform differently in different empirical situations, and ALASSO outperforms other methods when collinearity is low or moderate.
Article
Automation & Control Systems
Xiaoyu Jiang, Xiangyin Kong, Zhiqiang Ge
Summary: The curse of dimensionality refers to the problem of increased sparsity and computational complexity when dealing with high-dimensional data. In recent years, data augmentation technology has been introduced as an effective tool to solve the sparsity problem of high-dimensional industrial data. This paper systematically explores and discusses the necessity, feasibility, and effectiveness of augmented industrial data-driven modeling in the context of the curse of dimensionality and virtual big data. Then, the process of data augmentation modeling is analyzed, and the concept of data boosting augmentation is proposed.
IEEE-CAA JOURNAL OF AUTOMATICA SINICA
(2023)
Article
Cell Biology
Tianshi Lu, Seongoh Park, James Zhu, Yunguan Wang, Xiaowei Zhan, Xinlei Wang, Li Wang, Hao Zhu, Tao Wang
Summary: SClineager is a method that can infer accurate evolutionary lineages from scRNA-seq data, overcoming expressional drop-outs and applicable for single-cell sequencing studies. Genetics-based lineage tracing is not only feasible for tumor tissues, but also for non-tumor tissues research.
Article
Computer Science, Software Engineering
Daniel Alcaide, Jan Aerts
Summary: STAD is a parameter-free dimensionality reduction method that projects high-dimensional data into a graph, preserving approximate distances in the original high-dimensional space. This method can be used to explore and analyze data, highlighting potential traits in the data.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2021)
Article
Mathematics, Applied
Martin Hutzenthaler, Arnulf Jentzen, Thomas Kruse, Tuan Anh Nguyen
Summary: Backward stochastic differential equations (BSDEs) are extensively studied in stochastic analysis and computational stochastics. Nonlinear and high-dimensional BSDEs are common in real applications, but exact solutions are rarely attainable. Therefore, it is crucial to develop and analyze numerical approximation methods for solving these complex problems.
JOURNAL OF NUMERICAL MATHEMATICS
(2023)
Article
Multidisciplinary Sciences
Moein E. Samadi, Sandra Kiefer, Sebastian Johaness Fritsch, Johannes Bickenbach, Andreas Schuppert
Summary: Mechanistic/data-driven hybrid modeling is a crucial approach for situations where the mechanistic details are not well understood and purely data-driven models are too complex. By integrating first principles with data-driven approaches, hybrid modeling offers a feasible solution for data demand and extrapolation. This research introduces a learning strategy for tree-structured hybrid models in a binary classification task, specifically applied to predicting the vital status of COVID-19 patients.
Article
Chemistry, Multidisciplinary
Bardia Rafieian, Pedro Hermosilla, Pere-Pau Vazquez
Summary: In data science and visualization, dimensionality reduction techniques are widely used to explore large datasets by transforming high-dimensional data into reduced versions. Nonlinear approaches like t-SNE and UMAP have gained popularity in information visualization. This paper introduces a simple yet powerful manipulation for vector datasets that improves the results of dimensionality reduction algorithms across various scenarios and demonstrates improved clustering performance in data classification.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Caio Flexa, Walisson Gomes, Igor Moreira, Ronnie Alves, Claudomiro Sales
Summary: Dimensionality Reduction (DR) is important in understanding high-dimensional data, and the Polygonal Coordinate System (PCS) presented in this work offers an efficient geometric approach for this purpose. The study also introduces a new version of the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm using a PCS-based deterministic strategy, showcasing the efficiency of PCS in data embedding compared to other DR algorithms.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Statistics & Probability
Gabriel Chandler, Wolfgang Polonik
Summary: This paper proposes a method for extracting multiscale geometric features from a data cloud, and demonstrates its potential in various applications such as classification and anomaly detection. It also explores connections to other concepts such as random set theory, localized depth measures, and nonlinear dimension reduction.
ANNALS OF STATISTICS
(2021)
Article
Computer Science, Software Engineering
Laura Garrison, Juliane Mueller, Stefanie Schreiber, Steffen Oeltze-Jafra, Helwig Hauser, Stefan Bruckner
Summary: DimLift is a novel visual analysis method for creating and interacting with dimensional bundles. Dimensional bundles are expressive groups of dimensions that contribute similarly to the variance of a dataset. Through interactive exploration and reconstruction methods, users can lift interesting and subtle relationships to the surface.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2021)
Article
Computer Science, Artificial Intelligence
Qing Wang
Summary: The paper introduces a knowledge-based approach that combines case-based reasoning and operational research methodologies to solve repetitive combinatorial optimization problems, utilizing past experience to improve problem-solving efficiency, especially in cases where traditional schemes cannot solve the problem due to its large dimension.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Biochemical Research Methods
Jing Jiang, Junlin Xu, Yuansheng Liu, Bosheng Song, Xiulan Guo, Xiangxiang Zeng, Quan Zou
Summary: Single-cell RNA sequencing (scRNA-seq) is a revolutionary breakthrough for studying gene expressions at the individual cell level and understanding cell heterogeneity. However, scRNA-seq data are noisy, leading to challenges in dimensionality reduction and visualization. In this study, we propose an improved variational autoencoder model called DREAM, which combines different techniques to accurately analyze scRNA-seq data and identify cell types. Benchmarking comparisons show that DREAM outperforms current methods and can capture gene expression dynamics in human embryonic development.
BRIEFINGS IN BIOINFORMATICS
(2023)
Review
Biochemical Research Methods
Tallulah S. Andrews, Vladimir Yu Kiselev, Davis McCarthy, Martin Hemberg
Summary: Single-cell RNA sequencing (scRNA-seq) is a popular and powerful technology for profiling the whole transcriptome of individual cells, but analyzing the large volumes of data requires specialized statistical and computational methods. This article provides an overview of the computational workflow, common tasks and tools for addressing biological questions, as well as guidelines for best practices in computational analyses. It serves as a hands-on guide for experimentalists and an overview for bioinformaticians developing new computational methods.
Article
Automation & Control Systems
Jan Niklas Boehm, Philipp Berens, Dmitry Kobak
Summary: Neighbor embeddings are a family of methods for visualizing high-dimensional data sets using kNN graphs. By changing the balance between attractive and repulsive forces, a spectrum of embeddings with different trade-offs can be obtained. UMAP and ForceAtlas2 algorithms represent different levels of attraction on this spectrum.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Computer Science, Software Engineering
Eric Verner, Helen Petropoulos, Bradley Baker, Henry Jeremy Bockholt, Jill Fries, Anastasia Bohsali, Rajikha Raja, Duc Hoai Trinh, Vince Calhoun
Summary: BrainForge is a cloud-based neuroimaging analysis platform that allows users to archive and process data, as well as share results with colleagues. It addresses various challenges faced by researchers in neuroimaging data analysis, including software, reproducibility, computational resources, and data sharing.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2023)
Article
Neurosciences
Pujie Feng, Rongtao Jiang, Lijiang Wei, Vince D. Calhoun, Bin Jing, Haiyun Li, Jing Sui
Summary: This study investigates the impact of four confounding factors on individual trait prediction using resting-state functional connectivity (RSFC) data. The results suggest that the appropriate time series length and brain parcellation choice can improve prediction performance. Functional connectivity calculated by Pearson, Spearman, and Partial correlation achieves higher accuracy and lower time cost. Moreover, cognitive traits with larger variance among subjects can be better predicted.
Article
Cardiac & Cardiovascular Systems
Rongtao Jiang, Vince D. Calhoun, Stephanie Noble, Jing Sui, Qinghao Liang, Shile Qi, Dustin Scheinost
Summary: This study utilizes machine learning and functional connectivity to investigate the neurobiological correlates of blood pressure at an individual level. The results identify specific brain regions that are associated with blood pressure and provide evidence for meaningful neural representations of blood pressure in connectivity profiles.
CARDIOVASCULAR RESEARCH
(2023)
Article
Psychology, Developmental
Na Luo, Xiangsheng Luo, Suli Zheng, Dongren Yao, Min Zhao, Yue Cui, Yu Zhu, Vince D. Calhoun, Li Sun, Jing Sui
Summary: This study investigates the temporal and frequency abnormalities in ADHD and its subtypes using high-density EEG. The results show differences in the salience network and frequency power between ADHD patients and healthy controls. Subtype differences primarily exist in the visual network, with ADHD-C patients showing a more activated visual network. Furthermore, the support vector machine model achieves high accuracy in classifying ADHD and its subtypes.
EUROPEAN CHILD & ADOLESCENT PSYCHIATRY
(2023)
Article
Neurosciences
Md Abdur Rahaman, Jiayu Chen, Zening Fu, Noah Lewis, Armin Iraji, Theo G. M. van Erp, Vince D. Calhoun
Summary: Characterizing neuropsychiatric disorders is challenging, but combining structural and functional neuroimaging with genomic data in a multimodal classification framework can improve the classification of disorders and explore underlying neural and biological mechanisms. By developing neural networks for feature learning and implementing an adaptive control unit for fusion, we achieved high accuracy in schizophrenia prediction and identified critical neural features and genes/biological pathways associated with the disorder.
HUMAN BRAIN MAPPING
(2023)
Article
Neurosciences
Lei Wu, Vince Calhoun
Summary: The study of human brain connectivity provides insights into brain function and its relationship to behavior and cognition. Integrating structural connectivity and functional connectivity into a single framework is challenging. In this study, a new method called joint connectivity matrix independent component analysis (cmICA) is introduced, which allows for the integration of these two types of connectivity measurements using functional magnetic resonance imaging (MRI) and diffusion-weighted MRI data.
HUMAN BRAIN MAPPING
(2023)
Article
Biochemical Research Methods
Yuda Bi, Anees Abrol, Zening Fu, Jiayu Chen, Jingyu Liu, Vince Calhoun
Summary: Deep learning algorithms for predicting neuroimaging data have shown promise and outperform standard machine learning. This study focuses on using structural MRI data from the Adolescent Brain and Cognitive Development (ABCD) dataset to predict gender and identify gender-related changes in brain structure.
JOURNAL OF NEUROSCIENCE METHODS
(2023)
Article
Biochemistry & Molecular Biology
Constantinos Constantinides, Laura K. M. Han, Clara Alloza, Linda Antonella Antonucci, Celso Arango, Rosa Ayesa-Arriola, Nerisa Banaj, Alessandro Bertolino, Stefan Borgwardt, Jason Bruggemann, Juan Bustillo, Oleg Bykhovski, Vince Calhoun, Vaughan Carr, Stanley Catts, Young-Chul Chung, Benedicto Crespo-Facorro, Covadonga M. Diaz-Caneja, Gary Donohoe, Stefan Du Plessis, Jesse Edmond, Stefan Ehrlich, Robin Emsley, Lisa T. Eyler, Paola Fuentes-Claramonte, Foivos Georgiadis, Melissa Green, Amalia Guerrero-Pedraza, Minji Ha, Tim Hahn, Frans A. Henskens, Laurena Holleran, Stephanie Homan, Philipp Homan, Neda Jahanshad, Joost Janssen, Ellen Ji, Stefan Kaiser, Vasily Kaleda, Minah Kim, Woo-Sung Kim, Matthias Kirschner, Peter Kochunov, Yoo Bin Kwak, Jun Soo Kwon, Irina Lebedeva, Jingyu Liu, Patricia Mitchie, Stijn Michielse, David Mothersill, Bryan Mowry, Victor Ortiz-Garcia de la Foz, Christos Pantelis, Giulio Pergola, Fabrizio Piras, Edith Pomarol-Clotet, Adrian Preda, Yann Quide, Paul E. Rasser, Kelly Rootes-Murdy, Raymond Salvador, Marina Sangiuliano, Salvador Sarro, Ulrich Schall, Andre Schmidt, Rodney J. Scott, Pierluigi Selvaggi, Kang Sim, Antonin Skoch, Gianfranco Spalletta, Filip Spaniel, Sophia Thomopoulos, David Tomecek, Alexander S. Tomyshev, Diana Tordesillas-Gutierrez, Therese van Amelsvoort, Javier Vazquez-Bourgon, Daniela Vecchio, Aristotle Voineskos, Cynthia S. Weickert, Thomas Weickert, Paul M. Thompson, Lianne Schmaal, Theo G. M. van Erp, Jessica Turner, James H. Cole, Danai Dima, Esther Walton
Summary: Schizophrenia patients show evidence of advanced brain ageing, which is not associated with clinical characteristics.
MOLECULAR PSYCHIATRY
(2023)
Article
Computer Science, Interdisciplinary Applications
Harshvardhan Gazula, Kelly Rootes-Murdy, Bharath Holla, Sunitha Basodi, Zuo Zhang, Eric Verner, Ross Kelly, Pratima Murthy, Amit Chakrabarti, Debasish Basu, Subodh Bhagyalakshmi Nanjayya, Rajkumar Lenin Singh, Roshan Lourembam Singh, Kartik Kalyanram, Kamakshi Kartik, Kumaran Kalyanaraman, Krishnaveni Ghattu, Rebecca Kuriyan, Sunita Simon Kurpad, Gareth J. Barker, Rose Dawn Bharath, Sylvane Desrivieres, Meera Purushottam, Dimitri Papadopoulos Orfanos, Eesha Sharma, Matthew Hickman, Mireille Toledano, Nilakshi Vaidya, Tobias Banaschewski, Arun L. W. Bokde, Herta Flor, Antoine Grigis, Hugh Garavan, Penny Gowland, Andreas Heinz, Rudiger Bruhl, Jean-Luc Martinot, Marie-Laure Paillere Martinot, Eric Artiges, Frauke Nees, Tomas Paus, Luise Poustka, Juliane H. Frohner, Lauren Robinson, Michael N. Smolka, Henrik Walter, Jeanne Winterer, Robert Whelan, Jessica A. Turner, Anand D. Sarwate, Sergey M. Plis, Vivek Benegal, Gunter Schumann, Vince D. Calhoun
Summary: With the growth of decentralized/federated analysis approaches in neuroimaging, the opportunities to study brain disorders using data from multiple sites has grown multi-fold. One such initiative is the Neuromark, a fully automated spatially constrained independent component analysis (ICA) that is used to link brain network abnormalities among different datasets, studies, and disorders while leveraging subject-specific networks.
Article
Chemistry, Analytical
Hanlu Yang, Trung Vu, Qunfang Long, Vince Calhoun, Tuelay Adali
Summary: This study proposes a framework for subgroup identification of psychiatric patients using functional connectivity profiles obtained from fMRI data. The pipeline incorporates a data-driven method and constraint-based independent component analysis to identify meaningful subgroups with similar activation patterns in certain brain areas. The identified subgroups show significant group differences in multiple meaningful brain areas.
Article
Public, Environmental & Occupational Health
Erik Erhardt, Cristina Murray-Krezan, Lidia Regino, Daniel Perez, Elaine L. Bearer, Janet Page-Reeves
Summary: This article explores the association between depression and diabetes in a cohort of Latinx patients with diabetes from low-income households. The study found that culturally and contextually situated diabetes self-management programs, like the Chronic Care Model, had better outcomes for Latinx patients in terms of reducing depression and improving diabetes management.
SOCIAL SCIENCE & MEDICINE
(2023)
Article
Psychiatry
Zening Fu, Christopher C. Abbott, Jeremy Miller, Zhi-De Deng, Shawn M. McClintock, Mohammad S. E. Sendi, Jing Sui, Vince D. Calhoun
Summary: Electroconvulsive therapy (ECT) is effective for depression treatment, and its mechanism involves changing brain's functional organization through electrical current stimulation. This study investigated the relationship between whole-brain electric field (E-field), cerebro-cerebellar functional network connectivity (FNC), and clinical outcomes of ECT. The results showed that E-field influenced cognitive performance through cerebellum to middle occipital gyrus (MOG)/posterior cingulate cortex (PCC) FNC mediation, and had an effect on antidepressant outcomes through cerebellum to parietal lobule FNC mediation. Furthermore, larger E-field was associated with increased FNC between cerebellum and MOG and decreased FNC between cerebellum and PCC, which were linked with decreased cognitive performance.
TRANSLATIONAL PSYCHIATRY
(2023)
Article
Psychiatry
Kuaikuai Duan, Jiayu Chen, Vince D. D. Calhoun, Wenhao Jiang, Kelly Rootes-Murdy, Gido Schoenmacker, Rogers F. F. Silva, Barbara Franke, Jan K. K. Buitelaar, Martine Hoogman, Jaap Oosterlaan, Pieter J. J. Hoekstra, Dirk Heslenfeld, Catharina A. A. Hartman, Emma Sprooten, Alejandro Arias-Vasquez, Jessica A. A. Turner, Jingyu Liu
Summary: In this study, a genomic pattern underlying the gray matter variation in the frontal cortex related to working memory deficit in ADHD was revealed through a multivariate analysis. The identified genes are involved in modulating neuronal substrates underlying high-level cognition in ADHD, providing insights into the pathology of ADHD persistence.
TRANSLATIONAL PSYCHIATRY
(2023)
Article
Psychiatry
Lavinia Carmen Uscatescu, Martin Kronbichler, Sarah Said-Yurekli, Lisa Kronbichler, Vince Calhoun, Silvia Corbera, Morris Bell, Kevin Pelphrey, Godfrey Pearlson, Michal Assaf
Summary: Intrinsic neural timescales (INT) determine the duration of information storage in different brain areas. Previous studies have shown a hierarchy of INT from posterior to anterior regions in typically developed individuals (TD), as well as individuals with autism spectrum disorder (ASD) and schizophrenia (SZ), with shorter INT observed in both patient groups. This study aimed to replicate these group differences by comparing INT in TD, ASD, and SZ. The results partially replicated previous findings, showing reduced INT in the left lateral occipital gyrus and the right post-central gyrus in SZ compared to TD. Direct comparison between the patient groups also showed significantly reduced INT in SZ compared to ASD in these two areas. However, the previously reported correlations between INT and symptom severity were not replicated in this study. These findings help identify the brain areas that may play a crucial role in sensory peculiarities observed in ASD and SZ.
Article
Medicine, Research & Experimental
Allen J. Chang, Rebecca Roth, Eleni Bougioukli, Theodor Ruber, Simon S. Keller, Daniel L. Drane, Robert E. Gross, James Welsh, Anees Abrol, Vince Calhoun, Ioannis Karakis, Erik Kaestner, Bernd Weber, Carrie McDonald, Ezequiel Gleichgerrcht, Leonardo Bonilha
Summary: Chang et al. classified individuals with Temporal Lobe Epilepsy (TLE), Alzheimer's disease, and healthy controls using a convolutional neural network algorithm applied to magnetic resonance imaging (MRI) scans. They were able to distinguish people with TLE, including those without easily identifiable TLE-associated MRI features.
COMMUNICATIONS MEDICINE
(2023)
