A5076004172- Neural Networks and Applications
- Imbalanced Data Classification Techniques
- Advanced Memory and Neural Computing
- Fuzzy Logic and Control Systems
- Anomaly Detection Techniques and Applications
- Machine Learning and ELM
- Neural dynamics and brain function
- Face and Expression Recognition
- Electricity Theft Detection Techniques
- Model Reduction and Neural Networks
- Environmental Impact and Sustainability
- Microstructure and mechanical properties
- Machine Fault Diagnosis Techniques
- Shape Memory Alloy Transformations
- Probabilistic and Robust Engineering Design
- Corrosion Behavior and Inhibition
- Digital Media Forensic Detection
- Neural Networks and Reservoir Computing
- Machine Learning and Data Classification
- Industrial Vision Systems and Defect Detection
- Power Systems Fault Detection
- Explainable Artificial Intelligence (XAI)
- Energy, Environment, Economic Growth
- Gait Recognition and Analysis
- Fault Detection and Control Systems
Hunan University of Technology
2025
Dalian University of Technology
2012-2024
Shanghai Light Industry Research Institute
2024
Huazhong University of Science and Technology
2024
Ningbo University
2024
Hebei University of Technology
2008-2023
University of Science and Technology Beijing
2023
Liaoning University
2023
University of Shanghai for Science and Technology
2023
Chongqing Jiaotong University
2023
Data imbalance is a thorny issue in machine learning. SMOTE famous oversampling method of imbalanced However, it has some disadvantages such as sample overlapping, noise interference, and blindness neighbor selection. In order to address these problems, we present new method, OS-CCD, based on concept, the classification contribution degree. The degree determines number synthetic samples generated by for each positive sample. OS-CCD follows spatial distribution characteristics original class...
To improve the ability of deep learning model to handle imbalanced data, a fault diagnosis method based on improved gated convolutional neural network (IGCNN) is proposed. Firstly, an convolution layer proposed for feature extraction, with batch normalisation (BN) applied adjust data distribution and enhance generalisation performance model. Then, learned by multiple layers pooling fed fully connected type identification. Finally, label-distribution-aware margin (LDAM) loss function employed...
As a foundational clustering paradigm, Density Peak Clustering (DPC) partitions samples into clusters based on their density peaks, garnering widespread attention. However, traditional DPC methods usually focus high-density regions, neglecting representative peaks in relatively low-density areas, particularly datasets with varying densities and multiple peaks. Moreover, existing variants struggle to identify correctly high-dimensional spaces due the indistinct distance differences among...
Two backpropagation algorithms with momentum for feedforward neural networks a single hidden layer are considered. It is assumed that the training samples supplied to network in cyclic or an almost-cyclic fashion learning procedure, i.e., each cycle, sample of set fixed stochastic order respectively exactly once. A restart strategy adopted such coefficient zero at beginning cycle. Corresponding weak and strong convergence results then proved, indicating gradient error function goes weight...
Polysomnography (PSG) has been extensively studied for sleep staging, where stages are usually classified as wake, rapid-eye-movement (REM) sleep, or non-REM (NREM) (including light and deep sleep). Respiratory information proven to correlate with autonomic nervous activity that is related stages. For example, it known the breathing rate amplitude during NREM in particular steadier more regular compared periods of wakefulness can be influenced by body movements, conscious control, other...
Chaotic time series exist in nature, such as the field of meteorology or physics, with unpredictable features caused by their inherent high complexity and nonstationary motion. To improve prediction effect chaotic series, a hybrid method on basis empirical mode decomposition (EMD) neural networks (NN) is proposed. First, original decomposed into several intrinsic functions (IMFs) one residual EMD, whose components are divided high, medium low using runs test. The IMFs change dramatically,...
Induction motors (IMs) play an essential role in the field of various industrial applications. Long-time service and tough working situations make IMs become prone to a broken rotor bar (BRB) that is one major causes faults. Hence, continuous condition monitoring BRB faults demands computationally efficient accurate signal diagnosis technique. The advantage high reliability wide applicability fault based on vibration signature analysis results improved cyclic modulation spectrum (CMS), which...
Dropout and DropConnect are two techniques to facilitate the regularization of neural network models, having achieved state-of-the-art results in several benchmarks. In this paper, improve generalization capability spiking networks (SNNs), drop first applied SpikeProp learning algorithm resulting improved algorithms called SPDO (SpikeProp with Dropout) SPDC DropConnect). view that a higher membrane potential biological neuron implies probability activation, three adaptive algorithms,...
Abstract Purpose – With the function of reconstructing and promoting traditional industries in China, logistic service industry (LSI) still consumes a great deal energy. The purpose this paper is to empirically analyse relationship between energy consumption logistics its influential factors through this, identify most important factor give significant research afterwards. Design/methodology/approach Using quantitative analysis composition structure LSI, serious condition China's...
Induction motors (IMs) are widely used in many manufacturing processes and industrial applications. The harsh work environment, long-time enduring, overloads mean that it is subjected to broken rotor bar (BRB) faults. vibration signal of IMs with BRB faults consists the reliable modulation information for fault diagnosis. Cyclostationary analysis has been found be effective identifying extracting feature. estimators cyclic spectrum (CMS) fast spectral correlation (FSC) based on short-time...