Article
Construction & Building Technology
H. Yin, Z. Tang, C. Yang
Summary: This study proposed three models (BPNN, CCNN, and SVR) for predicting the electricity consumption of chillers in subway stations, and the results showed that these models could effectively predict the electricity consumption. CCNN and SVR exhibited higher accuracy compared to BPNN. By considering input characteristics, comparing different prediction models, and distinguishing different operating conditions, the prediction accuracy was improved effectively. This has significant implications for energy saving in subway stations.
JOURNAL OF BUILDING ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Zhijun Zhang, Guangqiang Chen, Song Yang
Summary: The article proposes a novel ensemble support vector recurrent neural network (E-SVRNN) framework for brain-computer interface applications, achieving more accurate and efficient EEG signal classification. Experimental results demonstrate its superior performance in accuracy and information transfer rate compared to most state-of-the-art algorithms.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Biochemistry & Molecular Biology
Nur Sakinah Ahmad Yasmin, Norhaliza Abdul Wahab, Fatimah Sham Ismail, Mu'azu Jibrin Musa, Mohd Hakim Ab Halim, Aznah Nor Anuar
Summary: The SVR models aim to predict the concentration of chemical oxygen demand in sequential batch reactors under high temperatures using the complex internal interaction between sludge characteristics and influent. The models were developed using a radial basis function algorithm with selected kernel parameters and optimized using particle swarm optimization and genetic algorithm. The results demonstrate the potential of SVR for simulating the complex aerobic granulation process.
Article
Geosciences, Multidisciplinary
Dechun Lu, Yiding Ma, Fanchao Kong, Caixia Guo, Jinbo Miao, Xiuli Du
Summary: This paper proposes a machine learning method based on heuristic optimization algorithms to predict the stratum displacement induced by earth pressure balanced shield tunneling. Support vector regression is used as the machine learning method, and heuristic intelligent optimization algorithms such as genetic algorithm, particle swarm optimization, grey wolf optimizer, and sparrow search algorithm are applied to optimize the hyperparameters of support vector regression model. The experimental results show that grey wolf optimizer and sparrow search algorithm are suitable methods for determining the hyperparameters.
Article
Computer Science, Artificial Intelligence
Quentin Klopfenstein, Samuel Vaiter
Summary: This paper investigates the addition of linear constraints to Support Vector Regression with a linear kernel, proving that the problem remains a semi-definite quadratic problem. A generalization of the Sequential Minimal Optimization algorithm is proposed to solve the optimization problem with linear constraints, showing convergence. Practical performance of this approach is demonstrated on simulated and real datasets, highlighting its usefulness compared to classical methods.
Article
Computer Science, Interdisciplinary Applications
Amir Feizi, Alireza Nazemi, Mohammad Reza Rabiei
Summary: This paper introduces a recurrent neural network to assist support vector machine learning in stochastic support vector regression, demonstrating its effectiveness in three illustrative examples.
ENGINEERING WITH COMPUTERS
(2022)
Article
Computer Science, Artificial Intelligence
Chenglong Zhang, Shifei Ding, Yuting Sun, Zichen Zhang
Summary: The study introduced an optimized support vector regression (SVR) model, which predicts the remaining useful life of bearings using features extracted by deep neural networks (DNN) and a novel multi-population fruit fly optimization algorithm (MPFOA).
APPLIED SOFT COMPUTING
(2021)
Review
Computer Science, Information Systems
Chieh-Huang Chen, Jung-Pin Lai, Yu-Ming Chang, Chi-Ju Lai, Ping-Feng Pai
Summary: Due to the rapid development in information technology, deep neural networks for regression (DNNR) have been widely used. The study aims to analyze the recent literature on DNNR optimization and investigate various platforms used for DNNR models. It provides sections for optimizing DNNR models, lists and analyzes elements of optimization in each section, and delivers insights and critical issues related to DNNR optimization. Simultaneous optimization of elements in each section can improve the performance of DNNR models. Possible directions for future study are also provided.
Article
Computer Science, Artificial Intelligence
Yafen Ye, Yuanhai Shao, Chunna Li, Xiangyu Hua, Yanru Guo
Summary: The paper introduces an online support vector quantile regression method, Online-SVQR, for dynamic time series with heavy-tailed noise. By using an incremental learning algorithm to update coefficients, Online-SVQR reflects dynamic information and outperforms traditional epsilon-support vector quantile regression in terms of sample selection ability and training speed.
APPLIED SOFT COMPUTING
(2021)
Article
Computer Science, Artificial Intelligence
Rakshitha Godahewa, Kasun Bandara, Geoffrey Webb, Slawek Smyl, Christoph Bergmeir
Summary: Ensembling techniques are used to improve the performance of Global Forecasting Models (GFM) and univariate models in heterogeneous datasets. A new clustered ensembles methodology is proposed to train multiple GFMs on different clusters of series, achieving higher accuracy than baseline models.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Thermodynamics
Hamed Khajavi, Amir Rastgoo
Summary: The growing global population has led to an increase in energy demand, particularly in the residential building sector where heating accounts for a significant portion of energy consumption. This study presents a machine learning model, specifically based on Support Vector Regression (SVR) combined with meta-heuristic algorithms, to accurately predict the heating energy consumption in residential buildings.
Article
Materials Science, Multidisciplinary
Zeljko Kanovic, Djordje Vukelic, Katica Simunovic, Miljana Prica, Tomislav Saric, Branko Tadic, Goran Simunovic
Summary: This research investigated ball burnishing of AISI 4130 alloy steel with high-stiffness tool and ceramic ball, modeling and predicting surface roughness during the process. The regression analysis model showed large prediction errors at low roughness values but was reliable for higher roughness values, while artificial neural network and support vector regression models exhibited excellent predictability across different input variables.
Article
Thermodynamics
Wei Cai, Xiaodong Wen, Chaoen Li, Jingjing Shao, Jianguo Xu
Summary: This study examines the impact of eight input factors on the heating and cooling loads of residential structures using the SVR algorithm. The results show that the SVR-AEO model performs the best in simulating residential buildings' heating and cooling loads.
Article
Computer Science, Artificial Intelligence
Juan Pablo Karmy, Julio Lopez, Sebastian Maldonado
Summary: A novel Support Vector Regression approach is proposed for dealing with hierarchical time series forecasting, pooling information across levels of hierarchy to prevent bottom-level series from deviating much with respect to the series at the upper levels. The proposed approach showed best performance compared with the state of the art on hierarchical time series forecasting using well-known benchmark datasets.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Multidisciplinary Sciences
Irfan Javid, Rozaida Ghazali, Irteza Syed, Muhammad Zulqarnain, Noor Aida Husaini
Summary: A stock market collapse refers to a situation where stock prices drop by more than 10% across major indexes. Predicting such crises is challenging, but a model combining Hybridized Feature Selection and deep learning algorithms can improve accuracy.