Article
Computer Science, Information Systems
Jing Bi, Yongze Lin, Quanxi Dong, Haitao Yuan, MengChu Zhou
Summary: The study presents a hybrid model based on a neural network and a filter for water quality time series prediction, with experimental results demonstrating better predictive performance compared to other similar models.
INFORMATION SCIENCES
(2021)
Article
Energy & Fuels
Meng Jiao, Dongqing Wang
Summary: This paper presents a SG-BiLSTM based method for SOC estimation of lithium batteries, which demonstrates advantages such as faster convergence speed, higher estimation accuracy, and strong robustness through experimental and simulation verification.
INTERNATIONAL JOURNAL OF ENERGY RESEARCH
(2021)
Article
Computer Science, Artificial Intelligence
Jing Bi, Shuang Li, Haitao Yuan, MengChu Zhou
Summary: Cloud computing providers face challenges in forecasting large-scale workload and resource time series. By using logarithmic operation, powerful filters, and deep learning methods, more accurate predictions can be achieved.
Article
Spectroscopy
Guosheng Zhang, He Hao, Yichen Wang, Ying Jiang, Jinhui Shi, Jing Yu, Xiaojuan Cui, Jingsong Li, Sheng Zhou, Benli Yu
Summary: An improved S-G filtering algorithm, combined with a deep learning network for real-time adjustment, effectively addresses the issue of blindly selecting filter parameters in digital signal processing. Compared to the MAF algorithm, the optimized S-G filtering algorithm demonstrates better performance in gas detection, with a sensitivity enhancement factor of 5.
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
(2021)
Article
Computer Science, Artificial Intelligence
Lei Chen, Weiwen Zhang, Haiming Ye
Summary: In this paper, a deep learning model SG-CBA is proposed for workload prediction, which combines SG filter, CNN, and BiLSTM with attention mechanism. The experimental results demonstrate that SG-CBA outperforms other alternatives in accurately predicting workload under various evaluation metrics.
APPLIED INTELLIGENCE
(2022)
Article
Chemistry, Analytical
Junfeng Ouyang, Changchun Chi
Summary: This paper proposes a method based on the Savitzky-Golay convolution smoothing long short-term memory neural network for predicting the electrical life of AC circuit breakers. By conducting a full lifespan test and utilizing feature extraction techniques, the proposed model achieves impressive accuracy of 97.4% in predicting the remaining electrical lifespan. This study demonstrates the feasibility of using time-series forecasting for predicting the residual electrical lifespan of electrical equipment and provides valuable insights for improving prediction methods.
Article
Thermodynamics
Fei Guo, Xiongwei Wu, Lili Liu, Jilei Ye, Tao Wang, Lijun Fu, Yuping Wu
Summary: In this paper, a prediction model for the state of health (SOH) and remaining useful life (RUL) of lithium batteries (LIBs) is developed by combining the Savitzky-Golay (SG) filter with gated recurrent unit (GRU) neural networks. Experiments and verification show that the proposed SG-GRU prediction model is effective for different applications, providing accurate prediction results under various charging strategies and different batteries. The model accurately tracks the nonlinear degradation trend of capacity during the whole cycle life, with a root mean square error of prediction controlled within 1%.
Article
Computer Science, Hardware & Architecture
Jing Bi, Haisen Ma, Haitao Yuan, Jia Zhang
Summary: This study proposes a hybrid prediction model called VAMBiG that integrates various techniques to address the challenges faced by cloud computing service providers in predicting large-scale workload and resource usage time series. The model utilizes signal decomposition, data pre-processing, bidirectional and grid LSTM networks, and a multi-head attention mechanism to achieve higher prediction accuracy. Experimental results demonstrate its superiority over several advanced prediction approaches.
IEEE TRANSACTIONS ON SUSTAINABLE COMPUTING
(2023)
Article
Geochemistry & Geophysics
Wei Liu, He Wang, Zhenzhu Xi, Rongqing Zhang
Summary: Despite the challenges in directly inverting MT field data using deep learning, this study proposes a new method that applies a multiwindow Savitzky-Golay filter to smooth the MT field measurements before network prediction. The smoothed data is then fed into a trained network for inversion. The use of layered resistivity models and the proposed MWSG filter enables smooth inversion. The efficiency of MT DL inversion is improved by incorporating Swin Transformer, and a physics-informed version is implemented to enhance generalization capability. The proposed method is demonstrated in both synthetic and field MT cases, showing improved adaptability and practicability for inverse problems in MT surveys.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Chemistry, Physical
Shaojie Zhang, Tao Chen, Fei Xiao, Rufeng Zhang
Summary: This paper proposes a degradation prediction model for proton exchange membrane fuel cells (PEMFC) based on a multi-reservoir echo state network with a mini reservoir. The model achieves high accuracy and robustness in the degradation prediction of PEMFC by optimizing the model parameters using the particle swarm optimization algorithm.
INTERNATIONAL JOURNAL OF HYDROGEN ENERGY
(2022)
Article
Environmental Sciences
Ruiqi Wang, Ying Qi, Qiang Zhang, Fei Wen
Summary: This paper presents a multi-step water quality prediction model for watersheds that combines Savitzky-Golay (SG) filter with Transformer optimized networks. It aims to improve the noise and sequence correlation issues in water quality data and solve the prediction lag problem by introducing the DILATE loss function. The experimental results demonstrate that the proposed model accurately predicts the shape, temporal positioning, and achieves the best accuracy in multi-step prediction tasks for multiple sites.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Chemistry, Analytical
Xi Wang, Chen Qian, Zhikai Zhao, Jiaming Li, Mingzhi Jiao
Summary: In recent years, the application of Deep Neural Networks has been developing in gas recognition. To improve the classification performance, various filtering methods are used to smooth filter the gas sensing response data and remove redundant information. The optimized Savitzky-Golay filtering algorithm is applied, and the gas sensing response data is encoded into two-dimensional sensing images using the Gramian Angular Summation Field (GASF) method. Data augmentation technology is used to enhance the classifier's robustness and generalization ability. With fine-tuning of the GoogLeNet neural network, the classification of four gases is achieved with high accuracy. The proposed method outperforms other networks like ResNet50, Alex-Net, and ResNet34 in terms of accuracy and sample processing times.
Article
Engineering, Multidisciplinary
Qingrong Wang, Xiaohong Chen, Changfeng Zhu, Wei Chai
Summary: Traffic volume forecasting is crucial for alleviating traffic congestion, however, the complex relationship between traffic data and outside factors complicates the problem. This study proposes a SGA-KGCN-LSTM model that integrates multiple techniques to address this issue. Experimental results demonstrate that the model achieves high forecasting accuracy compared to benchmark models and ablation experiments.
ENGINEERING LETTERS
(2023)
Article
Environmental Sciences
Kee-Won Seong, Jang Hyun Sung
Summary: The study proposed a method based on the SG filter to cope with oscillatory S-curves, which significantly reduces oscillation problems on the UH and IUH compared to traditional methods and has minimal impact on hydrograph properties. High levels of model performance criteria were demonstrated under storm data, suggesting potential scale issues for small watershed areas.
Article
Computer Science, Theory & Methods
K. Lalitha Devi, S. Valli
Summary: Resource management in a cloud setting is effectively achieved by predicting CPU and memory utilization using a hybrid ARIMA-ANN model. The combination of linear and nonlinear components in the prediction helps improve accuracy, and the introduction of a range of values reduces forecasting errors. OER and UER are used to cope with the error produced by over or under estimation of resource utilization.