Article
Engineering, Electrical & Electronic
Thanh T. Nguyen, Charles Soussen, Jerome Idier, El-Hadi Djermoune
Summary: This paper proposes an exact recovery analysis of greedy algorithms for non-negative sparse representations. The analysis shows that under certain conditions, the iterates of non-negative extensions coincide with those of orthogonal greedy algorithms. The sign preservation property of orthogonal greedy algorithms plays a crucial role in the analysis, and challenges in deriving improved analyses for correlated dictionaries are discussed.
Article
Mathematics
Luis Alberto Cantera-Cantera, Cristobal Vargas-Jarillo, Sergio Isai Palomino-Resendiz, Yair Lozano-Hernandez, Carlos Manuel Montelongo-Vazquez
Summary: This article introduces the classical curve-fitting problem and the related methods of least squares, total least squares, and orthogonal distances. The research shows that TLS and OD methods yield the same estimates when fitting a first-degree polynomial without an independent coefficient.
Article
Computer Science, Information Systems
Tapas Bhadra, Sanghamitra Bandyopadhyay
Summary: The paper proposes a novel supervised feature selection approach based on dense subgraph discovery. The algorithm proceeds in two phases to select features with maximal average class relevance, minimal average pairwise redundancy, and good discriminating power. Experimental results show the proposed approach is competitive with conventional and state-of-the-art algorithms in supervised feature selection.
INFORMATION SCIENCES
(2021)
Article
Engineering, Electrical & Electronic
Liyang Lu, Wenbo Xu, Yue Wang, Zhi Tian
Summary: This paper investigates the application of orthogonal least squares (OLS)-type algorithms in reconstructing sparse signals and provides exact recovery conditions for both noiseless and noisy scenarios. The theoretical analyses and simulation tests demonstrate the reliability and improvement of the algorithms.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2022)
Article
Chemistry, Analytical
James A. Jordan, Caelin P. Celani, Michael Ketterer, Barry K. Lavine, K. S. Booksh
Summary: This study demonstrates for the first time the use of supervised discretization to 'declutter' multivariate classification data in chemical sensor applications. The supervised discretization method is shown to be superior to the state-of-the-art EPO method in reducing within-class variance while retaining between-class variance. The results show that supervised discretization improves the performance of multivariate classification models.
Article
Acoustics
Nezih Topaloglu, Cevat V. Karadag
Summary: A linear regression based SDOF resonator parameter extraction method is proposed, which outperforms other methods using amplitude FRF.
JOURNAL OF SOUND AND VIBRATION
(2022)
Article
Food Science & Technology
Michele Ricci, Flavia Gasperi, Emanuela Betta, Leonardo Menghi, Isabella Endrizzi, Danny Cliceri, Pietro Franceschi, Eugenio Aprea
Summary: By analyzing two years of production data of Trentingrana cheese, it was found that the milk collection process has a significant impact on the content of organic acids, esters, and ketones, which are volatile organic compounds (VOCs) in the cheese.
LWT-FOOD SCIENCE AND TECHNOLOGY
(2023)
Article
Engineering, Electrical & Electronic
Yan-Chong Song, Fei-Yun Wu, Ru Peng
Summary: This article presents the orthogonal least squares (OLS) algorithm and its limitations in improving reconstruction accuracy. A neighborhood-based multiple orthogonal least squares (NMOLS) algorithm is proposed to address this issue.
Article
Engineering, Biomedical
Mingkan Shen, Peng Wen, Bo Song, Yan Li
Summary: This paper investigates the classification of alcoholic electroencephalogram (EEG) signals through whole brain connectivity analysis and deep learning methods. The study shows the accuracy of the proposed method by using 2D and 3D convolutional neural networks to classify alcoholic subjects and health control subjects.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Engineering, Chemical
Majed Aljunaid, Yang Tao, Hongbo Shi
Summary: The proposed MI-PLS method divides process variables into quality-related and quality-unrelated parts, providing lower computational load and more robust performance compared to Standard PLS.
Article
Mathematics
Luis Alberto Cantera-Cantera, Ruben Garrido, Luis Luna, Cristobal Vargas-Jarillo, Erick Asiain
Summary: This work presents the parameter identification of servo systems using the least squares of orthogonal distances method. The parameter identification problem is reformulated as data fitting to a plane, corresponding to a nonlinear minimization problem. Three servo system models with different numbers of parameters were experimentally identified using classic least squares and least squares of orthogonal distances. The results showed that the least squares of orthogonal distances method produced consistent estimates without the need for the classic persistency-of-excitation condition, and the parameter estimates obtained from this method achieved the best tracking performance when used in trajectory-tracking controller computation.
Article
Chemistry, Multidisciplinary
Yung-An Kao, Kun-Feng Wu
Summary: This paper proposes a low-complexity least-squares (LS) method to solve the leakage effect problem in DFT-based channel estimation. Compared to other methods, this method does not require knowledge of the statistical properties of the channel or the insertion of extra pilots, and has similar channel estimation efficiency to the LS method in simulation.
APPLIED SCIENCES-BASEL
(2022)
Article
Mathematics, Applied
Somayeh Mashayekhi
Summary: This paper introduces a new theoretical framework and numerical method to study the effect of heterogeneity in population genetics. It presents the fractional forward Kolmogorov equations as a solution to characterize the distribution of allele frequencies in a population. The paper also proposes a new numerical method for solving fractional partial differential equations and demonstrates its validity through illustrative examples.
COMMUNICATIONS IN NONLINEAR SCIENCE AND NUMERICAL SIMULATION
(2023)
Article
Engineering, Mechanical
Rohit Rana, Prerna Gaur, Vijyant Agarwal, Harish Parthasarathy
Summary: The research focuses on accurately estimating parameters from compressed temporal data in the presence of noise. The proposed method utilizes the properties of recursive wavelet domain to selectively store noise-free data coefficients, achieving data compression. The algorithm can be implemented on any scalable VLSI circuit and has been experimentally demonstrated on the Omni Bundle robot.
NONLINEAR DYNAMICS
(2022)
Article
Chemistry, Analytical
Hyeong Geun Jo, Beom Hoon Park, Do Yeong Joung, Jung Ki Jo, Jeong-Kyu Hoh, Won Young Choi, Kwan Kyu Park
Summary: The study introduces a wearable bladder scanner system that can continuously measure bladder volume in daily life. The system has shown similar measurement accuracy compared to commercial bladder imaging systems.
Article
Physics, Multidisciplinary
Lina Wang, Weining Xue, Yang Li, Meilin Luo, Jie Huang, Weigang Cui, Chao Huang
Article
Computer Science, Artificial Intelligence
Yang Li, Weigang Cui, Meilin Luo, Ke Li, Lina Wang
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS
(2018)
Article
Computer Science, Artificial Intelligence
Qinghua Wang, Hua-Liang Wei, Lina Wang, Song Xu
Summary: In this paper, a new time-varying modeling framework based on an autoregressive model is proposed to characterize and analyze EEG signals. The proposed method utilizes a multi-wavelet basis function expansion approach to approximate the TV parameters of the AR model, and an ultra-regularized orthogonal forward regression algorithm to refine the resulting expanded model. Experimental results show that the proposed TVAR-MWBF-UROFR method outperforms state-of-the-art classifiers in terms of accuracy, specificity, sensitivity, and robustness in seizure detection and classification tasks.
NEURAL COMPUTING & APPLICATIONS
(2021)