A5101790624- Machine Learning and ELM
- Energy Load and Power Forecasting
- Building Energy and Comfort Optimization
- Non-Destructive Testing Techniques
- Domain Adaptation and Few-Shot Learning
- Risk and Portfolio Optimization
- Innovative concrete reinforcement materials
- Stochastic processes and financial applications
- Brain Tumor Detection and Classification
- Ferroelectric and Negative Capacitance Devices
- Recycled Aggregate Concrete Performance
- Advanced Image Fusion Techniques
- Machine Fault Diagnosis Techniques
- Image Enhancement Techniques
- Structural Health Monitoring Techniques
- Concrete and Cement Materials Research
- Wind and Air Flow Studies
- Medical Image Segmentation Techniques
- Financial Markets and Investment Strategies
- Color Science and Applications
- Medical Imaging and Analysis
- Face and Expression Recognition
- Neural Networks and Applications
- Air Quality Monitoring and Forecasting
Shenzhen University
2020-2024
Xi'an University of Architecture and Technology
2021-2022
Southwestern University of Finance and Economics
2021
Shijiazhuang Tiedao University
2021
Guangdong Institute of Intelligent Manufacturing
2021
Dalian Maritime University
2017-2020
A general deep learning (DL) mechanism for a multiple hidden layer feed-forward neural network contains two parts, i.e., 1) an unsupervised greedy layer-wise training and 2) supervised fine-tuning which is usually iterative process. Although this has been demonstrated in many fields to be able significantly improve the generalization of network, there no clear evidence show one parts plays essential role improvement, resulting argument within DL community. Focusing on argument, paper...
Accurate forecasting of energy consumption in office buildings is great importance for optimal management and reduction building consumption. A variety combination models (FMs) have become the current research hotspots field forecasting. For problems large systematic errors poor generalization ability existing FMs, this paper proposes a dynamic residual model (FM) with approach. Firstly, support vector regression (SVR) selected as basic FM, SVR are forecasted by FM based on weights, forecast...
Segmentation on the left atrium from magnetic resonance imaging scans is a crucial task for clinical diagnosis of cardiac diseases. Automatic segmentation methods based deep learning have demonstrated their advantages in recent years, but they often suffer difficulty to process high uncertainty some regions. Focusing cognitive neural networks, we build an uncertainty-aware model anchor fuzzy sets this article. A entropy measure defined membership functions employed quantify induced by...
Energy consumption prediction can provide reliable data support for energy scheduling and optimization of office buildings. It is difficult traditional model to achieve stable accuracy robustness when mode complex sources are diverse. Based on such situation, this paper raised an approach containing the method comprehensive similar day ensemble learning. Firstly, historical was analyzed calculated obtain similarity degree meteorological features, time factor precursor. Next, entropy weight...