A5100607066- Blind Source Separation Techniques
- Neural Networks and Applications
- Advanced Computational Techniques and Applications
- Advanced Multi-Objective Optimization Algorithms
- Target Tracking and Data Fusion in Sensor Networks
- Control Systems and Identification
- Advanced Adaptive Filtering Techniques
- Traffic Prediction and Management Techniques
- Metaheuristic Optimization Algorithms Research
- Advanced machining processes and optimization
- Recommender Systems and Techniques
- Magnetic Properties and Applications
- Digital Marketing and Social Media
- E-commerce and Technology Innovations
- Chaos control and synchronization
- Building Energy and Comfort Optimization
- Simulation and Modeling Applications
Beijing University of Chemical Technology
2020
Shandong University of Science and Technology
2019
Shanghai Aerospace Automobile Electromechanical (China)
2018
Tzu Hui Institute of Technology
2010
University of Maryland, Baltimore County
2002
University System of Maryland
2002
Information Technology Laboratory
2002
Nature-inspired optimization is a modern technique in the past decades. Researchers report their successful applications various fields such as manufacturing, biomedical, and environmental engineering, while other researchers doubt its applicability. In this paper, we collect newly emerging nature-inspired algorithms proposed after 2008, present them unified way, implement them, evaluate on benchmark functions. Moreover, optimize behavioural parameters for these algorithms. Since it...
Information geometry of partial likelihood is constructed and used to derive the em-algorithm for learning parameters a conditional distribution model through information-theoretic projections. To construct coordinates information geometry, an expectation maximization (EM) framework described problem using Gaussian mixture probability model. It shown that information-geometric equivalent EM establish its convergence. The algorithm applied channel equalization by rapid convergence...
We introduce a unified statistical framework for real-time signal processing with neural networks by using recent extension of maximum likelihood (ML) estimation, partial (PL) estimation theory, which allows (i) dependent observations, and (ii) data only the information that is available at time processing. For general network conditional distribution model, we establish fundamental information-theoretic relationship PL obtain large sample properties case observations. consider applications...
We present a recurrent canonical piecewise linear (RCPL) network based on piecewise-linear (CPL) function and autoregressive moving average model, apply it to adaptive channel equalization. It is shown that neural with activation realizes an RCPL network. has several advantages. First, can make use of standard filtering techniques perform training tasks. Second, allows for efficient selection the partition boundaries corresponding appropriate complexity using CPL techniques. Third, being...
In order to solve the data sparsity problem existing in traditional collaborative filtering recommendation algorithm, a algorithm for scenic spots based on multi-dimensional feature clustering was proposed. Firstly, users are clustered and classified according vector. Then we determine category of target user. Building user-scenic spot score matrix, this basis, attention matrix is added. optimize similarity linearly combined with balance factor calculate between users. addition, threshold...
A recurrent canonical piecewise linear (RCPL) network is defined by combining the function with autoregressive moving average (ARMA) model such that an augmented input space partitioned into regions where ARMA used in each. Properties of RCPL are discussed. Particularly, it shown a contractive mapping and stable sense bounded output stability. By generalizing Donoho's minimum entropy deconvolution approach to nonlinear case, can achieve blind equalization. The applied both supervised...