A5016975682- Neural Networks and Applications
- Advanced Memory and Neural Computing
- Neural Networks and Reservoir Computing
- Domain Adaptation and Few-Shot Learning
- Industrial Technology and Control Systems
- Optical Network Technologies
- Advanced Sensor and Control Systems
- Machine Learning and ELM
- Sparse and Compressive Sensing Techniques
University of Jinan
2021-2022
Nitrogen oxide (NOx) is one of the important air pollutants in calcination process cement clinker. Accurate prediction NOx necessary for optimizing to reduce emissions. Based on EMD and multiple ESNs, this paper developed a model predict concentration kiln using history data. First, used decompose historical content time series into several subsequences. Then, each subsequence individually predicted by different ESN models, outputs models are integrated obtain final values. Finally,...
Echo state networks (ESNs) are a kind of special recurrent neural networks, which have super performance on time series predictions. However, ESNs with over large reservoirs may lead to the output collinearity problem. In this paper, regularized ESN an improved log penalty is proposed solve First, introduced quadratic loss function. Second, coordinate descent algorithm employed optimize function for sparse solution. Finally, model tested using two benchmark data sets. The experimental...
Echo State Networks (ESNs) are efficient recurrent neural networks (RNNs), which have attracted extensive attention due to their simple training processes and special reservoir structures. However, if the nonlinearity of network is improved, memory capability decreased. Therefore, this paper proposes a novel ESN model (MM-ESN) solve problem. We introduce linear (LMN) into based on idea relative separation for nonlinearity. improved while maintaining Experimental results Lorenz chaotic time...