A5100782028- Energy Load and Power Forecasting
- Market Dynamics and Volatility
- Neural Networks and Reservoir Computing
- Solar Radiation and Photovoltaics
- 3D Shape Modeling and Analysis
- Galectins and Cancer Biology
- Neural Networks and Applications
- Stock Market Forecasting Methods
- Energy, Environment, Economic Growth
- Grey System Theory Applications
- 3D Surveying and Cultural Heritage
- Model Reduction and Neural Networks
- Machine Learning and ELM
- Network Security and Intrusion Detection
- Image and Signal Denoising Methods
- Circular RNAs in diseases
- Evaluation and Optimization Models
- Wind Turbine Control Systems
- Regulation and Compliance Studies
- Amoebic Infections and Treatments
- Polymer composites and self-healing
- Geochemistry and Geochronology of Asian Mineral Deposits
- Immune Cell Function and Interaction
- Neural dynamics and brain function
- Photovoltaic System Optimization Techniques
Northwest Normal University
2021-2024
Fuzhou University
2009-2024
The First Affiliated Hospital, Sun Yat-sen University
2024
Zhongshan Hospital
2024
Fudan University
2024
Sun Yat-sen University
2024
Twist Bioscience (United States)
2024
Tianjin University
2024
Ningxia Medical University
2023
Renmin University of China
2022
Due to the strong randomness of wind speed, power generation is difficult integrate into grid. It very important predict speed reliably and accurately so that energy can be utilized effectively. In this study, obtain accurate prediction results, a combined VMD-D-ESN model based on variational mode decomposition (VMD), double-layer staged training echo state network (D-ESN) genetic algorithm (GA) optimization proposed. First, preprocesses original data with VMD then uses D-ESN each decomposed...
Accurate water level forecasting is essential for agricultural resources management, hydropower generation, flood control, drought relief, and watershed planning. A combined model (ICEEMDAN-VMD-WOA-ELM) proposed based on improved adaptive noise complete ensemble empirical mode decomposition (ICEEMDAN), variational (VMD), extreme learning machine (ELM) whale optimization algorithm (WOA). First, the historical data with high similarity to forecast date are extracted using a hierarchical...
Filled vulcanizates exhibit the nonlinear Payne effect under dynamic deformations at high strain amplitudes, while underlining mechanisms are still in dispute. Herein, polydimethylsiloxane (PDMS) networks prepared via end-linking of α,ω-vinyl-terminated PDMS chains by thiol-ene reaction, and influences silica on rheological responses investigated. The presence tends to improve crosslinking density alongside reinforcing network, itself does not contribute loss factor markedly. Furthermore,...
Reservoir computing is an effective model for learning and predicting nonlinear chaotic dynamical systems; however, there remains a challenge in achieving more dependable evolution such systems. Based on the foundation of Koopman operator theory, considering effectiveness sparse identification dynamics algorithm to construct candidate libraries application data, alternative reservoir method proposed, which creates linear Hilbert space system by including terms optimization process computing,...
Indoor point clouds often present significant challenges due to the complexity and variety of structures high object similarity. The local geometric structure helps model learn shape features objects at detail level, while global context provides overall scene semantics spatial relationship information between objects. To address these challenges, we propose a novel network architecture, PointMSGT, which includes multi-scale feature extraction (MSGFE) module Transformer (GT) module. MSGFE...
Nonferrous metals are the basic materials for national economic development. Accurate and robust price forecasting can effectively reduce investment costs bring greater benefits to enterprises. But violent fluctuation of nonferrous metal prices evolution irregular cycles make prediction difficult. A model based on improved complementary ensemble empirical mode (ICEEMDAN) decomposition, a Bayesian hyperparameter optimization gated recurrent neural network (GRU) an integrated autoregressive...
Immune checkpoint inhibitors (ICIs) targeting programmed cell death protein 1 (PD-1) and T immunoglobulin mucin domain-containing 3 (TIM-3) are beneficial to the resumption of anti-tumor immunity response hold extreme potential as efficient therapies for certain malignancies. However, ICIs with a single target exhibit poor overall rate in hepatocellular carcinoma (HCC) patients due complex pathological mechanisms HCC.
Due to the nonlinearity and high volatility of financial time series, hybrid forecasting systems combining linear nonlinear models can provide more precise performance than a single model. Therefore, this study proposes stock price model with error correction based on secondary decomposition. The modules data decomposition, prediction module an constitute overall framework proposed in paper. First, variational mode decomposition (VMD) decomposes original closing initial sequence. Second,...
Abstract Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) continues to evolve, escape disease 2019 therapeutics and vaccines, jeopardize public health. To combat SARS-CoV-2 antigenic escape, we developed a rapid, high-throughput pipeline discover monospecific VHH antibodies iteratively develop VHH-Fc-VHH bispecifics capable of neutralizing emerging variants. By panning single-domain phage libraries against ancestral or beta spike proteins, discovered high-affinity with unique...
Point clouds are essential 3D data representations utilized across various disciplines, often requiring point cloud completion methods to address inherent incompleteness. Existing like SnowflakeNet only consider local attention, lacking global information of the complete shape, and tend suffer from overfitting as model depth increases. To these issues, we introduced self-positioning point-based attention better capture contextual features designed a Channel Attention module for adaptive...
The Yellow River Basin is a major economic development area in China, and the high quality of basin still has great room for improvement. coordinated energy, environment, economy ecology one keys to watershed. Aiming at problem restricting promoting, this paper takes nine provinces as research object, based on data from 2004 2017, through establishing coupling coordination system ecology, using CRITIC method study characteristics each subsystem time space. results show that score phased...
Accurate prediction of wind power is crucial for the efficient operation and risk management farms. This paper introduces a deep learning model that integrates an Attention mechanism with convolutional neural network (CNN) gated recurrent unit (GRU) network. Addressing randomness, intermittency, volatility uncertainty speed, we first apply swarm decomposition (SWD) to preprocess original data into subsequences. Subsequently, CNN extracts spatial features, GRU identifies temporal...
Accurate recognition of mental workload is significant for optimizing the human-machine interaction and avoiding regrade task performance levels due to overloading or underloading workload. In past studies, use Electroencephalogram (EEG) signals has shown high in operators' levels, however, most studies were conducted using a single visual modality dual tasks. But real-world operational tasks, auditory-visual modalities tasks are commonly involved, there have been relatively few researches...
In the space clustering algorithm, because of choice value k and problem non-clear "elbow point" elbow method, this paper introduces logarithmic function determines initial center on basis properties exponential function, weight adjustment, bigotry term basic idea proposes an improved k-value selection algorithm. Combined with fully adaptive spectral global terrorist attack data are clustered. It effectively solves that is not clear outlier can be separated in process. The experimental...
Anomaly detection on attributed networks is crucial for many applications such as social security and fraud detection. Existing models utilize deep autoencoder to reconstruct structure attribute measure reconstruction errors obtain anomaly nodes. However, there are problems over-smoothing node's representation learning. In order solve these problems, in this paper, we propose a GAN-based Detection Attributed Networks (GANAN). The fake graph network generated by autoencoder. We use two...