Article
Computer Science, Information Systems
Guodong Li, Xvan Qin, He Liu, Kaiyuan Jiang, Aili Wang
Summary: In this paper, a new automatic modulation recognition algorithm is proposed, which extracts the constellation graph features of digital signals through clustering analysis and obtains the signal modulation mode using cascade SVM multi-classifiers. The method achieves a higher recognition accuracy at low SNR, indicating its effectiveness in noncooperation communication systems.
Article
Computer Science, Information Systems
Weichen Wu, Yitian Xu, Xinying Pang
Summary: A novel two-stage hybrid screening rule based on variational inequality and duality gap is proposed in this paper, which can accelerate the solving process of support vector machines by deleting more redundant samples while maintaining accuracy. This method also embeds shrinking technique into the fast iterative algorithm, further speeding up the solving process.
INFORMATION SCIENCES
(2021)
Article
Materials Science, Characterization & Testing
Junfei Nie, Xuelin Wen, Xuechen Niu, Yanwu Chu, Feng Chen, Weiliang Wang, Deng Zhang, Zhenlin Hu, Jinling Xiao, Lianbo Guo
Summary: A novel approach using LIBS combined with NCA and SVM was proposed to identify different colored plastics. The results showed that NCA-SVM achieved a higher average accuracy compared to SVM and PCA-SVM.
Article
Computer Science, Information Systems
Tee Yi Wen, Siti Armiza Mohd Aris
Summary: Support vector machine (SVM) algorithms are used to classify EEG signals for the detection of mental stress levels, with clustering methods employed to reduce subjective bias. The study shows that the activity of Beta frequency in the right prefrontal cortex significantly changes under stimuli, providing a highly accurate feature for predicting stress levels.
Article
Environmental Sciences
Guangxin Liu, Liguo Wang, Danfeng Liu, Lei Fei, Jinghui Yang
Summary: This article proposes a non-parallel SVM model, which improves the classification effect and generalization performance for hyperspectral images by adding an additional empirical risk minimization term and bias constraint.
Article
Computer Science, Artificial Intelligence
Yibang Ruan, Yanshan Xiao, Zhifeng Hao, Bo Liu
Summary: The article introduces a new method for support vector DML (CSV-DML) that uses SVM model instead of kNN model for data classification. By employing a nonlinear mapping to transform original instances into feature space and optimizing the CSV-DML model, it achieves better classification performance.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Chemistry, Multidisciplinary
Nur Izzati Mohd Talib, Nazatul Aini Abd Majid, Shahnorbanun Sahran
Summary: Predicting student academic success using data mining techniques is a major research issue in many academic fields. Monitoring and predicting student performance in higher education institutions is important for improving academic quality. This study aims to identify holistic features that form clusters and develop prediction models to predict student performance holistically. Both classification and clustering methods, SVM and K-means clustering, were used. Three clusters were identified based on learning program outcomes, representing low, average, and high performance students. The prediction model with new labels from the clusters achieved higher accuracy compared to using semester grades as labels.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Mohammed Alshutbi, Zhiyong Li, Moath Alrifaey, Masoud Ahmadipour, Muhammad Murtadha Othman
Summary: The decisions of experts and the evaluation of patient data play crucial roles in breast cancer analysis. Machine learning techniques can aid in quickly examining and diagnosing medical data, reducing potential errors caused by inexperienced decision-makers. This study proposes an intelligent cancer classification method that selects a feature subset and optimizes the parameters of the SVM classifier using the Jaya algorithm. The method is applied to accurately characterize a breast cancer dataset and compared with other classifiers, demonstrating its effectiveness.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Business, Finance
Kunpeng Yuan, Guotai Chi, Ying Zhou, Hailei Yin
Summary: Default prediction is crucial for banks and non-banking financial institutions to make lending decisions. This study proposes a two-stage default prediction model that integrates k-means clustering and support vector domain description (SVDD) to predict default, achieving a five-year default prediction ability with high accuracy. Key features for default forecasting in Chinese listed enterprises include retained earnings/total assets, financial expenses/gross revenue, and type of audit opinion.
RESEARCH IN INTERNATIONAL BUSINESS AND FINANCE
(2022)
Article
Computer Science, Information Systems
Ceren Atik, Recep Alp Kut, Reyat Yilmaz, Derya Birant
Summary: This paper proposes a novel method called support vector machine chains (SVMC) that involves chaining together multiple SVM classifiers in a special structure, decrementing one feature at each stage. The paper also introduces a new voting mechanism called tournament voting, where classifiers' outputs compete in groups and the winning class label of the final round is assigned as the prediction. Experimental results show that SVMC outperforms SVM in terms of accuracy and achieves a 6.88% improvement over state-of-the-art methods.
Article
Optics
Mengtian Yuan, Qingdong Zeng, Jie Wang, Wenxin Li, Guanghui Chen, Zitao Li, Yang Liu, Lianbo Guo, Xiangyou Li, Huaqing Yu
Summary: Improper disposal of steel waste is on the rise, leading to environmental pollution and resource wastage. A model combining PCA and SVM methods, along with FO-LIBS technology, can achieve rapid classification of special steel materials, contributing to resource conservation and online detection in the industrial field.
OPTICAL ENGINEERING
(2021)
Article
Chemistry, Analytical
Erik Kepes, Jakub Vrabel, Ondrej Adamovsky, Sara Stritezska, Pavlina Modlitbova, Pavel Porizka, Jozef Kaiser
Summary: This article presents four approaches to interpret support vector machines (SVMs) and investigates the classification task of 19 algal and cyanobacterial species. The study finds that different feature importance metrics provide complementary information, and identifies the SVM model's bias towards features with a large variance.
ANALYTICA CHIMICA ACTA
(2022)
Article
Mathematics
Guvenc Arslan, Ugur Madran, Duygu Soyoglu
Summary: In this paper, a novel classification approach is proposed by introducing a new clustering method as an intermediate step to discover the structure of a data set and reduce its size. Experimental results show that the proposed method performs comparably to standard support vector machines.
Article
Computer Science, Artificial Intelligence
Liming Liu, Maoxiang Chu, Rongfen Gong, Li Zhang
Summary: The improved nonparallel support vector machine (INPSVM) proposed in this article inherits the advantages of nonparallel support vector machine (NPSVM) while also offering incomparable benefits over twin support vector machine (TSVM). INPSVM effectively eliminates noise effects and achieves higher classification accuracy for both linear and nonlinear datasets compared to other algorithms. Experimental results demonstrate the superior efficiency, accuracy, and robustness of INPSVM.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Biotechnology & Applied Microbiology
Ke-Fan Wang, Jing An, Zhen Wei, Can Cui, Xiang-Hua Ma, Chao Ma, Han-Qiu Bao
Summary: In this paper, a novel imbalance classification method based on deep learning and fuzzy support vector machine, named DFSVM, is proposed. The method utilizes a deep neural network to obtain an embedding representation of the data and performs oversampling in the embedding space to address the data imbalance issue. Furthermore, a fuzzy support vector machine is used as the final classifier to improve the classification quality of minority classes.
FRONTIERS IN BIOENGINEERING AND BIOTECHNOLOGY
(2022)
Article
Engineering, Environmental
Shunchun Yao, Lifeng Zhang, Yeming Zhu, Junye Wu, Zhimin Lu, Jidong Lu
Article
Energy & Fuels
Shunchun Yao, Xiayang Yao, Lifeng Zhang, Huaiqing Qin, Ziyu Yu, Xiaoxuan Chen, Zhimin Lu, Jidong Lu
Article
Energy & Fuels
Hao Luo, Lukasz Niedzwiecki, Amit Arora, Krzysztof Moscicki, Halina Pawlak-Kruczek, Krystian Krochmalny, Marcin Baranowski, Mayank Tiwari, Anshul Sharma, Tanuj Sharma, Zhimin Lu
Article
Thermodynamics
Hao Luo, Zhimin Lu, Peter Arendt Jensen, Peter Glarborg, Weigang Lin, Kim Dam-Johansen, Hao Wu
Summary: The impact of gasification reactions on biomass char conversion was studied through single particle experiments and modeling. It was found that char oxidation is limited by mass transfer, while char gasification is controlled by both mass transfer and gasification kinetics. Sensitivity analysis revealed that CO oxidation and gasification kinetics significantly influence char conversion time.
PROCEEDINGS OF THE COMBUSTION INSTITUTE
(2021)
Article
Spectroscopy
Shunchun Yao, Ziyu Yu, Shuixiu Xu, Xiayang Yao, Huaiqing Qin, Zhimin Lu, Jidong Lu
Summary: The research shows that reducing spherical aberration in the focusing configuration improves laser-particle coupling and enhances plasma stability for more accurate analysis.
SPECTROCHIMICA ACTA PART B-ATOMIC SPECTROSCOPY
(2021)
Article
Energy & Fuels
Xin Li, Zhimin Lu, Jinzheng Chen, Xiaoxuan Chen, Yuan Jiang, Jie Jian, Shunchun Yao
Summary: The study demonstrates that oxidative torrefaction (OT) affects the combustion process of wood spheres, increasing char yield and combustion time while forming a more cross-linked aromatic structure that hinders volatile release, resulting in denser and heavier char particles. The findings suggest that mass yield serves as a good indicator for both OT and traditional inert torrefaction processes.
Article
Green & Sustainable Science & Technology
Zhimin Lu, Xiaoxuan Chen, Yuan Jiang, Xin Li, Jinzheng Chen, Yuesheng Li, Weiye Lu, Jidong Lu, Shunchun Yao
Summary: The study demonstrated the feasibility of fuel property analysis of biomass fuel using LIBS technique. PLS models were built to evaluate the performance on wood pellets, showing satisfactory results with high accuracy in determining fuel properties without further sample preparation. The R2 values of the calibration models were all above 0.98, indicating the effectiveness of using LIBS for direct monitoring of wood pellet quality.
Article
Instruments & Instrumentation
Huaiqing Qin, Ziyu Yu, Zhimin Lu, Zhuliang Yu, Shunchun Yao
Summary: This paper proposes a method to improve the performance of laser-induced breakdown spectroscopy (LIBS) quantitative analysis by using data uncertainty. The method preserves some data uncertainty in the calibration data matrix to optimize the calibration model, resulting in better accuracy and robustness compared to conventional methods.
APPLIED SPECTROSCOPY
(2022)
Article
Spectroscopy
Shunchun Yao, Huaiqing Qin, Shuixiu Xu, Ziyu Yu, Zhimin Lu, Zhe Wang
Summary: A synergistic method based on LIBS and NIRS is proposed for optimal coal proximate analysis. The ash and moisture content are analyzed separately, and then the fusion data from LIBS and NIRS are used to analyze the calorific value and volatile matter. The fixed carbon content is calculated by subtracting the mass percentages of ash, moisture, and volatile matter. The results show that the proposed method achieves good quantitative performance for coal proximate analysis.
ATOMIC SPECTROSCOPY
(2022)
Article
Spectroscopy
Shao Guo-dong, Li Zheng-hui, Guo Song-jie, Zou Li-chang, Deng Yao, Lu Zhi-min, Yao Shun-chun
Summary: Tunable diode laser absorption spectroscopy technology is widely used in various fields due to its strong selectivity, high sensitivity, high accuracy, and non-invasive measurement capabilities. The direct absorption technology, which does not require precalibration, fits gas concentrations directly with improved accuracy compared to traditional methods. The method based on the gradient descent approach shows promise in achieving more accurate gas concentration measurements with reduced errors.
SPECTROSCOPY AND SPECTRAL ANALYSIS
(2021)
Article
Chemistry, Analytical
Yuan Jiang, Zhimin Lu, Xiaoxuan Chen, Ziyu Yu, Huaiqing Qin, Jinzheng Chen, Jidong Lu, Shunchun Yao
Summary: Rapid analysis of fuel properties is crucial for solid biomass utilization, and the combination of LIBS technology with KPLS method offers improved accuracy. The study demonstrated better quantitative analysis performance through various data normalization methods.
ANALYTICAL METHODS
(2021)
Article
Chemistry, Analytical
Ziyu Yu, Shunchun Yao, Yuan Jiang, Weize Chen, Shuixiu Xu, Huaiqing Qin, Zhimin Lu, Jidong Lu
Summary: This paper investigates the matrix effect in coal analysis using laser-induced breakdown spectroscopy (LIBS), comparing the effect in coal particle flow with that in coal pellets. The study demonstrates that analyzing coal particle flow with LIBS can reduce the matrix effect and improve the stability of analytical results.
JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY
(2021)
Article
Engineering, Electrical & Electronic
Deng Yao, Tang Wen, Li Zhenghui, Zhong Shangwen, Zou Lichang, Lu Zhimin, Lin Jian, Yao Shunchun
Summary: The paper proposed a new gas-concentration inversion method based on direct absorption peak value calibration. By analyzing the theoretical changes of absorption peak value with concentration and temperature in a numerical model, and establishing an actual concentration calibration model and temperature correction curve, accurate measurement of high concentrations of CO2 was achieved.
LASER & OPTOELECTRONICS PROGRESS
(2021)
Article
Spectroscopy
Zou Li-chang, Huang Jun, Li Zheng-hui, Deng Yao, Shao Guo-dong, Ruan Zhen, Lu Zhi-min, Yao Shun-chun
Summary: Tunable Diode Laser Absorption Spectroscopy (TDLAS) is widely used for combustion diagnosis, measuring trace gases, and industrial process control. By utilizing wavelength modulation spectroscopy's second-harmonic (2f) detection technology, gas sensing can be achieved effectively. Research has shown that through specific laser selection and signal processing methods, the adverse effects of 2f background signal drift on gas concentration inversion results can be eliminated, thereby improving the sensitivity, accuracy, and stability of the TDLAS-WMS online detection system.
SPECTROSCOPY AND SPECTRAL ANALYSIS
(2021)
Article
Spectroscopy
Shunchun Yao, Huaiqing Qin, Qi Wang, Zhimin Lu, Xiayang Yao, Ziyu Yu, Xiaoxuan Chen, Lifeng Zhang, Jidong Lu
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
(2020)
