A5102923053- Neural Networks and Applications
- Face and Expression Recognition
- Matrix Theory and Algorithms
- Numerical methods for differential equations
- Machine Learning and ELM
- Advanced Clustering Algorithms Research
- Image Retrieval and Classification Techniques
- Time Series Analysis and Forecasting
- Text and Document Classification Technologies
- Stock Market Forecasting Methods
- Metaheuristic Optimization Algorithms Research
- Blind Source Separation Techniques
- Anomaly Detection Techniques and Applications
- Data Stream Mining Techniques
- Soil Carbon and Nitrogen Dynamics
- Energy Load and Power Forecasting
- Domain Adaptation and Few-Shot Learning
- Natural Language Processing Techniques
- Advanced Numerical Methods in Computational Mathematics
- Visual Attention and Saliency Detection
- Electromagnetic Simulation and Numerical Methods
- Artificial Immune Systems Applications
- Statistical and numerical algorithms
- Music and Audio Processing
- Evolutionary Algorithms and Applications
Chinese Academy of Sciences
2021-2024
Institute of Applied Ecology
2021-2024
Affiliated Hospital of Southwest Medical University
2024
University of Chinese Academy of Sciences
2021-2023
Dongzhimen Hospital Affiliated to Beijing University of Chinese Medicine
2022
Beijing University of Chinese Medicine
2022
Nanjing University
2009-2018
Nanjing University of Science and Technology
2013-2016
Kunming University of Science and Technology
2016
Ministry of Industry and Trade
2015
Abstract Mental disorders are the leading contributors to globally nonfatal burden of disease. This study was aimed estimate mental in Asian countries. Based on GBD 2019, prevalence and disability-adjusted life years (DALYs) rates with 95% uncertainty intervals (UI) were estimated Predictions for future 8 selected countries, ranks correlations Sociodemographic Index (SDI) also estimated. During past 3 decades, while number DALYs increased from 43.9 million (95% UI: 32.5–57.2) 69.0...
In recent years, neural networks is increasingly adopted in the prediction of exchange rate. However, most them predict a specific number, which can not help speculators too much because small gap between predicted values and actual will lead to disastrous consequences. our study, purpose present model forecast fluctuation range rate by combining Fuzzy Granulation with Continuous-valued Deep Belief Networks (CDBN), concept "Stop Loss" introduced for making environment profit strategy close...
Forecasting exchange rates is an important financial problem which has received much attention. Nowadays, neural network become one of the effective tools in this research field. In paper, we propose use a deep belief (DBN) to tackle rate forecasting problem. A DBN applied predict both British Pound/US dollar and Indian rupee/US our experiments. We six evaluation criteria evaluate its performance. also compare method feedforward (FFNN), state-of-the-art for with networks. Experiments...
Dongkuan Xu, Ian En-Hsu Yen, Jinxi Zhao, Zhibin Xiao. Proceedings of the 2021 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies. 2021.
The proposed perception evolution network (PEN) is a biologically inspired neural model for unsupervised learning and online incremental learning. It able to automatically learn suitable prototypes from data in an way, it does not require the predefined prototype number or similarity threshold. Meanwhile, being more advanced than existing model, PEN permits emergence of new dimension field network. When introduced, integrate dimensional sensory inputs with learned prototypes, i.e., are...
The imbalance problem exists in P300 EEG data sets because potential are collected under the condition of Oddball experimental paradigm. Hence, a detection method, namely RUSBagging SVMs, is proposed this paper to solve and make an improvement. This algorithm re-samples at first generate rebalanced training set one round iteration trains SVM classifier based on set. Next, classifiers integrated final decision. In integration several classifiers, information that lost under-sampling process...
Time series classification (TSC) problem is important due to the pervasiveness of time data. Shapelet provides a mechanism for by its ability measure local shape similarity. However, shapelets need be searched from massive sub-sequences. To address this problem, paper proposes novel shapelet learning method classification. The proposed uses self-organizing incremental neural network learn candidates. learned candidates reduce greatly in quantity and improve much quality. After that, an...
The aim of this paper is to develop a unified special extended Nyström tree (SEN-tree) theory which provides theoretical framework for the order conditions multidimensional Runge–Kutta–Nyström (ERKN) methods proposed by X. Wu et al. (Wu al., 2010). new SEN complete and consistent, has overcome drawback bi-coloured in H. Yang al.'s work (Yang 2009) where two "branch sets" have be constructed true solutions numerical solutions, respectively.