A5100714793- Control Systems and Identification
- Neural Networks and Applications
- Fault Detection and Control Systems
- Structural Health Monitoring Techniques
- Sparse and Compressive Sensing Techniques
- Fuzzy Logic and Control Systems
- Blind Source Separation Techniques
- Advanced Adaptive Filtering Techniques
- Face and Expression Recognition
- Advanced Algorithms and Applications
- Spectroscopy and Chemometric Analyses
- Statistical Methods and Inference
- Advanced Wireless Communication Techniques
- Gaussian Processes and Bayesian Inference
- Machine Learning and ELM
- Criminal Law and Evidence
- Advanced Power Amplifier Design
- Probabilistic and Robust Engineering Design
- PAPR reduction in OFDM
- Imbalanced Data Classification Techniques
- Advanced Computational Techniques and Applications
- Model Reduction and Neural Networks
- Congenital heart defects research
- Remote Sensing and Land Use
- Tensor decomposition and applications
Northeast Normal University
2022-2025
Sir Run Run Shaw Hospital
2024-2025
Zhejiang University
2024-2025
China University of Geosciences
2025
University of Reading
2015-2024
Institute of Engineering Thermophysics
2024
Chinese Academy of Sciences
2024
Xiyuan Hospital
2022-2024
China Academy of Chinese Medical Sciences
2024
Zhejiang Chinese Medical University
2022-2024
The paper introduces an efficient construction algorithm for obtaining sparse linear-in-the-weights regression models based on approach of directly optimizing model generalization capability. This is achieved by utilizing the delete-1 cross validation concept and associated leave-one-out test error also known as predicted residual sums squares (PRESS) statistic, without resorting to any other data set evaluation in process. Computational efficiency ensured using orthogonal forward...
Many kernel classifier construction algorithms adopt classification accuracy as performance metrics in model evaluation. Moreover, equal weighting is often applied to each data sample parameter estimation. These modeling practices become problematic if the sets are imbalanced. We present a algorithm using orthogonal forward selection (OFS) order optimize generalization for imbalanced two-class sets. This identification based on new regularized weighted least squares (ROWLS) estimator and...
In the ICEBERG project at UC Berkeley, we are developing an Internet-based integration of telephony and data services spanning diverse access networks. Our primary goals extensibility, scalability, robustness, personalized communication. We leverage Internet's low cost entry for service creation, provision, deployment, integration. present our solutions to signaling, easy resource reservation, admission control, billing, security in network architecture.
An improved mixture of probabilistic principal component analysis (PPCA) has been introduced for nonlinear data-driven process monitoring in this paper. To realize purpose, the technique a analysers is utilized to establish model underlying with local PPCA models, where novel composite statistic proposed based on integration two statistics modified PPCA-based fault detection approach. Besides, weighted mean aforementioned utilised as metrics detect potential abnormalities. The virtues...
Subspace clustering groups a set of samples from union several linear subspaces into clusters, so that the in same cluster are drawn subspace. In majority existing work on subspace clustering, clusters built based feature information, while sample correlations their original spatial structure simply ignored. Besides, high-dimensional vector contains noisy/redundant and time complexity grows exponentially with number dimensions. To address these issues, we propose tensor low-rank...
The note proposes an efficient nonlinear identification algorithm by combining a locally regularized orthogonal least squares (LROLS) model selection with D-optimality experimental design. proposed aims to achieve maximized robustness and sparsity via two effective complementary approaches. LROLS method alone is capable of producing very parsimonious excellent generalization performance. design criterion further enhances the efficiency robustness. An added advantage that user only needs...
We propose a unified data modeling approach that is equally applicable to supervised regression and classification applications, as well unsupervised probability density function estimation. A particle swarm optimization (PSO) aided orthogonal forward (OFR) algorithm based on leave-one-out (LOO) criteria developed construct parsimonious radial basis (RBF) networks with tunable nodes. Each stage of the construction process determines center vector diagonal covariance matrix one RBF node by...
Tensor clustering is an important tool that exploits intrinsically rich structures in real-world multiarray or datasets. Often dealing with those datasets, standard practice to use subspace based on vectorizing data. However, vectorization of tensorial data does not exploit complete structure information. In this paper, we propose a algorithm without adopting any process. Our approach novel heterogeneous Tucker decomposition model taking into account cluster membership We new alternates...
This paper introduces a novel sparse dynamic inner principal component analysis (SDiPCA) based monitoring for multimode processes. Different from traditional algorithms, model is updated sequential modes by memorizing the significant features of existing modes. By adopting concept intelligent synapses in continual learning, loss quadratic term introduced to penalize changes mode–relevant parameters, where modified synaptic intelligence (MSI) proposed estimate parameter importance. Thus,...
Atmospheric fine particulate matter (PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> ) poses significant risks to both environmental and human health, highlighting the need for regional estimations spatiotemporal analyses. While most studies have focused on large-scale areas, such as global or national levels, fewer addressed PM at urban level. This study analyzed monitoring data from ground stations in Wuhan, collected between July...
A framework for portfolio allocation based on multiple hypotheses prediction using structured ensemble models is presented. Portfolio optimization formulated as an learning problem, where each predictor focuses a specific asset or hypothesis. The weights are determined by optimizing the ensemble's parameters, equal-weighted target, serving canonical basis hypotheses. Diversity in among predictors parametrically controlled, and their predictions form input model. proposed methodology...
Objective:The aim of this study was to investigate the emotional experience patients with thin endometrium (TE) who have repeatedly cancelled their cycles due unsuitability for embryo implantation during endometrial preparation phase freeze-thaw transfer (FET).The overall is improve management strategies and quality life these patients.Methods: A descriptive phenomenological methodology utilized conduct in-depth, semi-structured interviews ten diagnosed TE had experienced repeated FET...
Abstract Maps of regional morbidity and mortality rates are useful tools in determining spatial patterns disease. Combined with sociodemographic census information, they also permit assessment environmental justice; that is, whether certain subgroups suffer disproportionately from diseases or other adverse effects harmful exposures. Bayes empirical methods have proven smoothing crude maps disease risk, eliminating the instability estimates low-population areas while maintaining geographic...
A novel particle swarm optimisation (PSO) tuned radial basis function (RBF) network model is proposed for identification of non-linear systems. At each stage orthogonal forward regression (OFR) construction process, PSO adopted to tune one RBF unit's centre vector and diagonal covariance matrix by minimising the leave-one-out (LOO) mean square error (MSE). This aided OFR automatically determines how many tunable nodes are sufficient modelling. Compared with the-state-of-the-art local...
This paper describes a novel on-line learning approach for radial basis function (RBF) neural network. Based on an RBF network with individually tunable nodes and fixed small model size, the weight vector is adjusted using multi-innovation recursive least square algorithm on-line. When residual error of becomes large despite adaptation, insignificant node little contribution to overall system replaced by new node. Structural parameters are optimized proposed fast algorithms in order...
We investigated the prevalence of Ureaplasma spp. in semen samples infertile men Shanghai, China and evaluated correlation between sperm parameters (seminal volume, concentration, progressive motility non-progressive) secretary function these infectious populations.Semens were collected from 540 260 fertile control group shanghai, subjected to standard bacterial culture. Positive isolates further tested by PCR detect biovars serotypes Sperm seminological variabilities analyzed...