Article
Computer Science, Artificial Intelligence
Taiyong Li, Zijie Qian, Wu Deng, Duzhong Zhang, Huihui Lu, Shuheng Wang
Summary: The paper introduces a novel approach VMD-RSBL that combines VMD and RSBL for forecasting crude oil prices. By decomposing the crude oil price series into components and predicting each component individually using predictors built on random samples and random lags, the final forecasted prices are obtained by aggregating the predictions of all components. Extensive experiments show that VMD-RSBL outperforms many state-of-the-art schemes in terms of several evaluation indicators.
APPLIED SOFT COMPUTING
(2021)
Article
Economics
Michal Rubaszek
Summary: The study suggests that long-term forecasts should utilize the mean reversion of real oil prices and models should not replicate the high volatility of oil prices observed in samples. The DSGE model performs better in forecasting real oil prices compared to random walk or vector autoregression models.
INTERNATIONAL JOURNAL OF FORECASTING
(2021)
Article
Energy & Fuels
Laiba Sultan Dar, Muhammad Aamir, Zardad Khan, Muhammad Bilal, Nattakan Boonsatit, Anuwat Jirawattanapanit
Summary: This study proposes a novel hybrid method for accurately predicting the volatility of crude oil prices. By splitting EEMD IMFs into stochastic and deterministic components and testing with ARIMA and FFNN models, the proposed hybrid model performs well in terms of performance. Simulation results demonstrate the robustness of the method under different quantities.
FRONTIERS IN ENERGY RESEARCH
(2022)
Article
Computer Science, Artificial Intelligence
Guancen Lin, Aijing Lin, Jianing Cao
Summary: Stock time series forecasting is a crucial purpose for academic researchers. In this paper, a modified modeling procedure combining EEMD and MKNN-TSPI methods is proposed to enhance prediction accuracy. Experimental results show that the proposed EEMD-MKNN-TSPI model outperforms other models, indicating its effectiveness in stock market prediction.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Engineering, Electrical & Electronic
Chunlian Xia, Jing Huang
Summary: This paper introduces the application of empirical mode decomposition method in inflation forecasting, which decomposes the time series into eigenmode functions with different time scales to predict inflation rate more accurately. The predicted results show that the method's predicted values are relatively close to the actual values, indicating a better prediction effect of the EEMD model.
JOURNAL OF SENSORS
(2022)
Article
Environmental Studies
Krzysztof Drachal
Summary: This research aims to forecast real crude oil prices using Time-Varying Vector Autoregression models, with model averaging and selection schemes to address variable uncertainty. The study finds that the Vector Autoregression approach leads to more accurate forecasts compared to Time-Varying Regression or Dynamic Model Averaging, with a model combination scheme of multiple Vector Autoregression models outperforming a single approach.
Article
Business, Finance
Kunliang Xu, Weiqing Wang
Summary: A reliable crude oil price forecast is crucial for market pricing. This study incorporates a rolling window into two prevalent EEMD-based modeling paradigms to improve accuracy. The results show that EEMD plays a weak role in improving crude oil price forecasts when only the in-sample set is preprocessed, but the rolling EEMD-denoising model has an advantage for long-term forecasting.
INTERNATIONAL REVIEW OF FINANCIAL ANALYSIS
(2023)
Article
Environmental Studies
He Jiang, Weiqiang Hu, Ling Xiao, Yao Dong
Summary: This study combines a decomposition-ensemble approach with a sentiment analysis to accurately forecast crude oil prices. The approach involves obtaining a cumulative sentiment score sequence through sentiment analysis of news headlines data, decomposing historical crude oil prices using an adaptive signal decomposition method, optimizing the hyperparameters of gated recurrent units using a seagull optimization algorithm, and integrating the forecasting results of each component using multiple linear regression. The empirical results confirm the effectiveness of this approach in forecasting crude oil prices and also analyze the impact of black-swan events on price fluctuations.
Article
Engineering, Electrical & Electronic
Chunlian Xia, Jing Huang
Summary: This paper investigates the factors affecting the performance of Bayesian models in inflation forecasting and introduces the Empirical Mode Decomposition method (EEMD) into inflation forecasting, resulting in constructive conclusions.
JOURNAL OF SENSORS
(2022)
Article
Thermodynamics
Guohui Li, Shibo Yin, Hong Yang
Summary: This paper proposes a novel crude oil prices forecasting model based on secondary decomposition with improved complementary ensemble empirical mode decomposition with adaptive noise, state space correlation entropy, improved variational mode decomposition by tunicate swarm algorithm, and improved kernel based extreme learning machine by artificial gorilla troops optimizer. The model decomposes the nonstationary, nonlinear and highly complex time series into regular sub-series through secondary decomposition. Then, parameter-optimized extreme learning machine is utilized to forecast all the sub-series, and the forecast values are reconstructed to obtain accurate crude oil prices forecasts.
Article
Computer Science, Artificial Intelligence
Xiangjun Cai, Dagang Li
Summary: This paper presents a new decomposition mechanism based on learned decomposition mapping. By using a neural network to learn the relationship between original time series and decomposed results, the repetitive computation overhead during rolling decomposition is relieved. Additionally, extended mapping and partial decomposition methods are proposed to alleviate boundary effects on prediction performance. Comparative studies demonstrate that the proposed method outperforms existing RDEMs in terms of operation speed and prediction accuracy.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Engineering, Electrical & Electronic
Tingting Zhang, Zhenpeng Tang, Junchuan Wu, Xiaoxu Du, Kaijie Chen
Summary: Research on forecasting electricity prices is crucial for market participants. Due to the nonlinearity and high volatility of electricity prices, forecasting the price series is challenging. This study introduces a two-layer decomposition technique and an optimized hybrid model to improve prediction accuracy and demonstrates the advantages of the proposed model through empirical analysis.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Business
Ranran Li, Yucai Hu, Jiani Heng, Xueli Chen
Summary: This study explores a novel multiscale hybrid paradigm to estimate crude oil prices, which decomposes the price into simple models to improve accuracy in forecasting complex time series.
TECHNOLOGICAL FORECASTING AND SOCIAL CHANGE
(2021)
Article
Computer Science, Artificial Intelligence
Yan Fang, Wenyan Wang, Pengcheng Wu, Yunfan Zhao
Summary: This study proposes a new hybrid forecasting approach, FinBERT-VMD-Att-BiGRU, which effectively predicts the price of crude oil. The approach combines FinBERT for extracting news information, VMD for decomposing price series, an attention mechanism for weighting input features, and BiGRU for price forecasting. Experimental results demonstrate that the proposed model outperforms other benchmarks in forecasting performance, accuracy, and trading strategy.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Green & Sustainable Science & Technology
Fanhua Yu, Huibowen Hao, Qingliang Li
Summary: The study proposes DL models combined with Ensemble Empirical Mode Decomposition to better capture the complex spatiotemporal relationship of soil temperature, resulting in improved performance. EEMD-Conv3d demonstrates the best performance among experimental models, with higher R2 values, lower RMSE, and closer alignment between predicted and observed ST. The results suggest that EEMD-Conv3D is a more effective method for predicting spatiotemporal soil temperature.
Article
Thermodynamics
Yingjie Song, Daqing Wu, Wu Deng, Xiao-Zhi Gao, Taiyong Li, Bin Zhang, Yuangang Li
Summary: The proposed MPPCEDE algorithm optimizes parameters of PV models and enhances solar energy conversion efficiency through the reverse learning mechanism, multi-population parallel control strategy, and co-evolutionary mutation strategy, demonstrating higher accuracy, reliability, and fast convergence speed compared to other methods.
ENERGY CONVERSION AND MANAGEMENT
(2021)
Article
Engineering, Electrical & Electronic
Taiyong Li, Jiayi Shi, Duzhong Zhang
Summary: A joint permutation and diffusion method for color image encryption is proposed, utilizing a 4D hyperchaotic system to generate sequences and introducing plain image information for enhanced security. Experimental results show its capability to resist various attack types.
JOURNAL OF ELECTRONIC IMAGING
(2021)
Article
Physics, Multidisciplinary
Duzhong Zhang, Lexing Chen, Taiyong Li
Summary: The paper introduces a novel image encryption method based on a six-dimensional hyper-chaotic encryption scheme, utilizing three-dimensional transformed Zigzag diffusion and RNA operation, showing high resistance against known attacks and superior performance in protecting digital visual information.
Article
Physics, Multidisciplinary
Taiyong Li, Duzhong Zhang
Summary: This paper proposes a novel hyperchaotic image encryption scheme MBPD, which enhances image security through multiple bit permutation and diffusion. Extensive experiments show that MBPD can effectively resist different types of attacks and outperforms other popular encryption methods.
Article
Computer Science, Artificial Intelligence
Taiyong Li, Jiayi Shi, Wu Deng, Zhenda Hu
Summary: This paper proposes a novel particle swarm optimization algorithm called PPSO, which utilizes a pyramid structure and competitive-cooperative strategies to update particle information. Extensive experiments demonstrate that PPSO outperforms other algorithms in terms of accuracy and convergence speed, indicating its potential in numerical optimization.
APPLIED SOFT COMPUTING
(2022)
Article
Physics, Multidisciplinary
Zhiyi Wang, Mingcheng Zhou, Boji Liu, Taiyong Li
Summary: This paper proposes a novel scheme using Transformer for feature extraction in image steganography, which is shown to outperform other deep-learning models in terms of feature extraction. Additionally, an image encryption algorithm with good attributes for image security is also proposed, further enhancing the performance of the steganography scheme.
Article
Physics, Multidisciplinary
Jiaxuan Xu, Jiang Wu, Taiyong Li, Yang Nan
Summary: This paper proposes a divergence-based locally weighted ensemble clustering with dictionary learning (DLWECDL) method to address the issues of microcluster weights and sample-cluster relationship in ensemble clustering. Experimental results demonstrate the great potential of this method in improving clustering accuracy.
Article
Computer Science, Information Systems
Duzhong Zhang, Xiancheng Wen, Chao Yan, Taiyong Li
Summary: This paper presents a novel joint RNA-level permutation and substitution (JRPS) based image encryption algorithm, which utilizes a six-dimensional hyper-chaotic system to generate pseudo-random sequences and transforms plaintext images into RNA codon sequences according to RNA rules. Running two rounds of joint RNA-level permutation and substitution on the RNA codon sequence yields a cipher image. Simulation results demonstrate that the proposed algorithm can withstand various attacks.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Software Engineering
Jiayi Shi, Taiyong Li, Jiaxuan Xu
Summary: This paper proposes a recursive lightweight CNN approach (PPNets) that achieves significant improvement in image denoising, with fewer model parameters compared to traditional models and state-of-the-art CNN models.
Article
Computer Science, Information Systems
Duzhong Zhang, Lexing Chen, Taiyong Li
Summary: This paper presents a new image encryption algorithm called HCLRNA, which is based on a hyper-chaotic system, three-dimensional orthogonal Latin cube transformation, and RNA diffusion. It consists of three main steps: generating chaotic matrices using a 6D hyper-chaotic system, scrambling the plaintext image using 3D orthogonal Latin cube transformation, and diffusing the scrambled pixel values using RNA codons. Experimental results show that HCLRNA meets the requirements of different evaluation indicators, effectively resists common attacks, and performs significantly better in resisting differential attacks compared to other studies.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Qingxiao Zheng, Lingfeng Wang, Jin He, Taiyong Li
Summary: With the expansion of services in cloud manufacturing, service level agreements (SLAs) are increasingly used by cloud manufacturers to ensure business processing cooperation. However, consensus algorithms in Blockchain as a Service (BaaS) systems often overlook the importance of SLAs. To address this issue, a KNN-based consensus algorithm is proposed that classifies transactions based on their priority. The enhanced consensus algorithm improves the satisfaction of SLAs in BaaS systems, allowing cloud service providers to provide better services to cloud service consumers.
Article
Computer Science, Information Systems
Panjie Wang, Jiang Wu, Yuan Wei, Taiyong Li
Summary: This study proposes a hybrid ensemble learning algorithm, CEEMD-MultiRocket, which combines Complementary Ensemble Empirical Mode Decomposition (CEEMD) with an improved MultiRocket for accurate time series classification. The method decomposes the raw time series into IMFs and a residue using CEEMD, and selects the decomposed sub-series based on their classification accuracy compared to the raw time series. The improved MultiRocket is then applied to the selected sub-series and the first-order difference of the raw time series to generate the final classification results.
Article
Physics, Multidisciplinary
Wei Fan, Taiyong Li, Jianan Wu, Jiang Wu
Summary: This paper proposes a novel scheme for color image encryption based on eight-base DNA-level permutation and diffusion. The experimental results demonstrate the excellent performance of the proposed scheme in color image encryption and its resistance to various attacks.
Editorial Material
Computer Science, Information Systems
Taiyong Li, Wu Deng, Jiang Wu
Article
Computer Science, Artificial Intelligence
Jiaxuan Xu, Taiyong Li, Duzhong Zhang, Jiang Wu
Summary: This paper proposes a novel scheme called FSEC to improve the performance of ensemble clustering by integrating both global and local structural information into a learning framework. Experimental results demonstrate that FSEC outperforms other state-of-the-art methods of ensemble clustering.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)