Article
Acoustics
Zhipeng Dong, Yucheng Liu, Jianshe Kang, Shaohui Zhang
Summary: This paper proposes a new method called stochastic learning algorithm (SL) to address the nonlinear problems in industrial big data. The SL method reduces the dimension of high-dimensional data, enhances the clustering influence of samples, and denoises the data to improve classification accuracy while reducing computational burden.
SHOCK AND VIBRATION
(2022)
Article
Engineering, Chemical
Amin Mahdavi-Meymand, Wojciech Sulisz
Summary: In this study, ARIA models were developed to enhance the prediction of boiling point rise in a multi-stage flash desalination system. The ARIA models showed greater accuracy and increased prediction efficiency compared to regular models. The ARIA-ANFIS model performed the best, reducing the error in RF predictions by 69.66%.
Article
Materials Science, Multidisciplinary
Yi Wu, Xueqin Chen, Dongqi Jiang
Summary: In this study, a deep forest (DF) model was developed to predict central deflection. It showed stronger learning and generalization abilities compared to other deep learning algorithms. The DF model with custom backend settings outperformed the highly encapsulated model with sklearn as the backend in terms of predictive performance and computation time.
Article
Computer Science, Information Systems
Dinghao Yang, Wei Gao, Ge Li, Hui Yuan, Junhui Hou, Sam Kwong
Summary: In this paper, an efficient point cloud classification method based on manifold learning is proposed. The method embeds point cloud features using manifold learning algorithms to consider the geometric continuity on the surface. Experimental results show that the proposed method outperforms existing methods and achieves better classification accuracy.
ACM TRANSACTIONS ON MULTIMEDIA COMPUTING COMMUNICATIONS AND APPLICATIONS
(2023)
Article
Environmental Sciences
Lifei Wei, Kun Wang, Qikai Lu, Yajing Liang, Haibo Li, Zhengxiang Wang, Run Wang, Liqin Cao
Summary: The study proposes a fine classification method for extracting crop types from airborne hyperspectral images, utilizing spatial information through multi-feature fusion and deep learning. Features such as morphological profiles, GLCM texture, and endmember abundance are used, along with the deep neural network with conditional random field model, to achieve satisfactory classification performance.
Article
Biochemistry & Molecular Biology
Leqi Tian, Wenbin Wu, Tianwei Yu
Summary: Random Forest (RF) is a popular machine learning method for classification and regression tasks, and it performs well under low sample size situations. However, there are issues with gene selection using RF as the important genes are usually scattered on the gene network, which conflicts with the biological assumption of functional consistency. To address this issue, we propose the Graph Random Forest (GRF) method, which incorporates external topological information to identify highly connected important features. The algorithm achieves equivalent classification accuracy to RF while selecting interpretable feature sub-graphs.
Article
Biochemistry & Molecular Biology
Albert Roethel, Piotr Bilinski, Takao Ishikawa
Summary: In this study, a new deep neural network architecture called BioS2Net is proposed for extracting sequential and structural information of biomolecules. The performance of BioS2Net is evaluated on two protein fold classification datasets, demonstrating its effectiveness and reliability in protein fold recognition.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Chunhui Wu, Wenjuan Li
Summary: This study investigates feature selection methods and introduces an ensemble of Neural Networks and Random Forest to enhance intrusion detection performance. The experimental results show that compared to similar approaches, this method can better identify important and relevant features.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Environmental Sciences
Kongyang Zhu, Chao Shen, Chen Tang, Yixi Zhou, Chengyong He, Zhenghong Zuo
Summary: The study developed a virtual screening procedure to identify potential AhR ligands, with significantly higher precision and recall rates than previous methods, indicating that supervised machine learning techniques can improve virtual screening performance. Of 77 pesticides screened, 77 were identified as potential AhR ligands, with 12 being reported as AhR agonists for the first time.
Article
Biology
Timothy Szocinski, Duc Duy Nguyen, Guo-Wei Wei
Summary: This study introduces a new neural network model, AweGNN, which can automatically extract features of complex biomolecular data, overcoming the obstacle of manual parametrization. By constructing multi-task and single-task models, AweGNN demonstrates state-of-the-art performance in molecular property predictions.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Plant Sciences
Ximeng Cheng, Ali Doosthosseini, Julian Kunkel
Summary: In forestry studies, this research aims to improve the interpretability of deep learning models and enhance their performance through guided training using expertise. The experiments demonstrate improved models based on explanation assessment and the automatic generation of expertise in the form of annotation matrix. The study emphasizes the importance of model interpretation and improvement based on expertise in deep learning research.
FRONTIERS IN PLANT SCIENCE
(2022)
Article
Environmental Sciences
Hao Fei, Zehua Fan, Chengkun Wang, Nannan Zhang, Tao Wang, Rengu Chen, Tiecheng Bai
Summary: This study proposes a county-scale cotton mapping method based on multiple features and random forest. By selecting spectral features, vegetation indices, and texture features, and exploring the contribution of texture features to cotton classification accuracy, the study improves the accuracy and efficiency of cotton classification.
Article
Computer Science, Information Systems
Haibin Liao, Dianhua Wang, Ping Fan, Ling Ding
Summary: The study tackles the challenges of automated facial expression recognition by introducing a conditional random forest architecture and a deep multi-instance learning model. Experimental results demonstrate that the proposed method performs well on public datasets and exhibits good robustness in complex environments.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Mathematics
Muyi Chen, Daling Wang, Shi Feng, Yifei Zhang
Summary: In this paper, the representation learning problem is studied from an out-of-distribution (OoD) perspective to identify the factors affecting representation quality. The concept of out-of-feature subspace (OoFS) noise is introduced, and its reduction is proven to be beneficial for better representation. A novel data-dependent regularizer is proposed to reduce noise in representations and achieve better performance in multiple tasks.
Article
Biochemistry & Molecular Biology
Hasan Zulfiqar, Shi-Shi Yuan, Qin-Lai Huang, Zi-Jie Sun, Fu-Ying Dao, Xiao-Long Yu, Hao Lin
Summary: This study developed a computational model to discriminate cyclin proteins from non-cyclin proteins with high accuracy. By encoding and optimizing protein sequences with seven features, and training a gradient boost decision tree classifier, the model achieved better results than previous studies on the same data.
COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL
(2021)
Article
Biochemical Research Methods
Teng Fei, Tianwei Yu
Review
Biochemical Research Methods
Kenong Su, Tianwei Yu, Hao Wu
Summary: Cell clustering is a crucial task in single-cell RNA sequencing (scRNA-seq) data analysis, with feature selection playing a key role in improving clustering accuracy. The study evaluates the impact of feature selection on cell clustering accuracy and introduces a new algorithm named FEAture SelecTion (FEAST) for selecting more representative features. Applying FEAST to 12 public scRNA-seq datasets demonstrates a significant improvement in clustering accuracy when combined with existing clustering tools.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Virology
Samantha McInally, Kristin Wall, Tianwei Yu, Rabindra Tirouvanziam, William Kilembe, Jill Gilmour, Susan A. Allen, Eric Hunter
Summary: The study showed that individuals who later became HIV-1 infected had significantly higher baseline levels of multiple inflammatory cytokines/chemokines compared to individuals who remained HIV-negative. Specific levels of certain biomarkers were identified as significant predictors of later HIV acquisition, indicating a potential link between inflammation and immune activation with increased risk of HIV infection.
Article
Immunology
Nadine G. Rouphael, Lilin Lai, Sonia Tandon, Michele Paine McCullough, Yunchuan Kong, Sarah Kabbani, Muktha S. Natrajan, Yongxian Xu, Yerun Zhu, Dongli Wang, Jesse O'Shea, Amy Sherman, Tianwei Yu, Sebastien Henry, Devin McAllister, Daniel Stadlbauer, Surender Khurana, Hana Golding, Florian Krammer, Mark J. Mulligan, Mark R. Prausnitz
Summary: The study found that inactivated influenza virus vaccination through dissolvable microneedle patches (MNPs) produces humoral and cellular immune responses that are similar or greater than traditional intramuscular (IM) vaccination. MNPs induced higher neuraminidase inhibition (NAI) titers for all three influenza virus strains tested and stimulated a larger percentage of circulating T follicular helper cells.
Article
Environmental Sciences
Song Tang, Tiantian Li, Jianlong Fang, Renjie Chen, Yu'e Cha, Yanwen Wang, Mu Zhu, Yi Zhang, Yuanyuan Chen, Yanjun Du, Tianwei Yu, David C. Thompson, Krystal J. Godri Pollitt, Vasilis Vasiliou, John S. Ji, Haidong Kan, Junfeng Jim Zhang, Xiaoming Shi
Summary: The exposome is a novel research paradigm that comprehensively considers the complex interactions between exogenous exposures, endogenous exposures, and modifiable factors in humans. By exploring the association between individual airborne exposure and adverse health outcomes, utilizing advanced monitoring techniques and biological sample analysis, the exposome approach can reveal the mechanisms underlying the impact of environmental exposures on human health.
ENVIRONMENT INTERNATIONAL
(2021)
Article
Engineering, Environmental
Ziyin Tang, Jeremy A. Sarnat, Rodney J. Weber, Armistead G. Russell, Xiaoyue Zhang, Zhenjiang Li, Tianwei Yu, Dean P. Jones, Donghai Liang
Summary: The study found that particulate oxidative potential may be a key parameter for particulate matter toxicity. By examining the biological changes and underlying molecular mechanisms associated with particulate oxidative potential, the study identified leukotriene metabolism and galactose metabolism in plasma, and vitamin E metabolism and leukotriene metabolism in saliva as top pathways associated with FPMOP. The study also observed different patterns of perturbed pathways for water-soluble and -insoluble FPMOP, and identified five metabolites directly associated with FPMOP. These findings suggest that FPMOP may be a more sensitive and health-relevant measure for understanding the causes related to PM2.5 exposures.
ENVIRONMENTAL SCIENCE & TECHNOLOGY
(2022)
Article
Mathematical & Computational Biology
Zhuxuan Jin, Jian Kang, Tianwei Yu
Summary: Jointly analyzing transcriptomic data and existing biological networks leads to more robust and informative feature selection results and a better understanding of biological mechanisms. A new Bayesian node classification framework is proposed to handle missing values and improve classification accuracy while reducing bias in estimating gene effects. This method outperforms existing approaches in comprehensive simulation studies and analysis of real-world genomic data.
STATISTICS IN MEDICINE
(2022)
Article
Biochemical Research Methods
Tianwei Yu
Summary: This study introduces a deep learning-based approach, AIME, for extracting data representation for integrative analysis of omics data. The method can adjust for confounding factors, achieve informative data embedding, and identify related feature pairs between two data types.
PLOS COMPUTATIONAL BIOLOGY
(2022)
Article
Biochemical Research Methods
Leqi Tian, Zhenjiang Li, Guoxuan Ma, Xiaoyue Zhang, Ziyin Tang, Siheng Wang, Jian Kang, Donghai Liang, Tianwei Yu
Summary: This article introduces an innovative R/Bioconductor package for pathway enrichment testing of untargeted metabolomics data. The package addresses the matching uncertainty between data features and metabolites, and allows for the simultaneous analysis of positive and negative ion mode LC/MS data.
Article
Engineering, Environmental
Zhenjiang Li, Jeremy A. Sarnat, Ken H. Liu, Robert B. Hood, Che-Jung Chang, Xin Hu, ViLinh Tran, Roby Greenwald, Howard H. Chang, Armistead Russell, Tianwei Yu, Dean P. Jones, Donghai Liang
Summary: Saliva may serve as an alternative biospecimen to blood in evaluating the association between traffic-related air pollution and biological responses.
ENVIRONMENTAL SCIENCE & TECHNOLOGY
(2022)
Article
Biology
Junning Feng, Yong Liang, Tianwei Yu
Summary: Personalized treatment of complex diseases relies on combined medication, but unexpected drug-drug interactions (DDIs) can lead to adverse effects. This study proposes a multimodal graph-agnostic neural network model for predicting drug-drug interaction events. The model demonstrates competitive performance on prediction tasks, particularly in predicting DDI types for new drugs, and outperforms existing methods in terms of accuracy, F1 score, precision, and recall.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Neurosciences
Zhuxuan Jin, Jian Kang, Tianwei Yu
Summary: This study proposes a Bayesian hierarchical model to investigate the shape and intensity of brain activation regions, and develops efficient posterior computation algorithms. The results demonstrate the significant application value of this model in Alzheimer's disease research.
FRONTIERS IN NEUROSCIENCE
(2023)
Article
Endocrinology & Metabolism
Victoria M. Pak, Katherine Russell, Zhenzhen Shi, Qiang Zhang, John Cox, Karan Uppal, Tianwei Yu, Vicki Hertzberg, Ken Liu, Octavian C. Ioachimescu, Nancy Collop, Donald L. Bliwise, Nancy G. Kutner, Ann Rogers, Sandra B. Dunbar
Summary: There is a difference in 24-hour MAP between sleepy and non-sleepy participants with newly diagnosed OSA, and sphinganine is significantly associated with MAP in non-sleepy patients with OSA.
Article
Multidisciplinary Sciences
Tengjiao Zhang, Yichi Xu, Kaoru Imai, Teng Fei, Guilin Wang, Bo Dong, Tianwei Yu, Yutaka Satou, Weiyang Shi, Zhirong Bao
Article
Mathematics, Interdisciplinary Applications
Qingpo Cai, Jian Kang, Tianwei Yu