Article
Automation & Control Systems
Assen L. Dontchev, Ilya V. Kolmanovsky, Trung B. Tran
Summary: This article considers the problem of best smoothing in a strip and proposes a numerical algorithm for solving it. Numerical results are reported to demonstrate the effectiveness of the proposed approach.
OPTIMAL CONTROL APPLICATIONS & METHODS
(2022)
Article
Computer Science, Artificial Intelligence
Jaiprakash Nagar, Sanjay Kumar Chaturvedi, Sieteng Soh, Abhilash Singh
Summary: This study proposes a machine learning approach based on the generalized regression neural network (GRNN) to predict the k-coverage performance of wireless multihop networks (WMNs) placed in a rectangular region. The proposed approach achieves better prediction accuracy and lower computational time complexity compared to existing benchmark algorithms in both scenarios with and without boundary effects (BEs).
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Physics, Multidisciplinary
Rui Pan, Zhanfeng Wang, Yaohua Wu
Summary: This article proposes a method for studying function-on-function regression models by constructing a tensor product space of reproducing kernel Hilbert spaces and using a model selection approach. It allows for estimating the functional function with functional covariate inputs and detecting interaction effects among the functional covariates.
Article
Statistics & Probability
Wensheng Guo, Mengying You, Jialin Yi, Michel A. Pontari, J. Richard Landis
Summary: By clustering UCPPS patients into homogeneous subgroups and utilizing longitudinal data modeling, we are able to identify different trajectories of the disease and associate them with clinical outcomes and baseline predictors. The proposed method shows promising performance in simulation studies and identifies four distinct subgroups of UCPPS patients, each with varying disease progression patterns.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2022)
Article
Automation & Control Systems
Hui Jin, Guido Montufar
Summary: This article investigates the training of wide neural networks using gradient descent and the corresponding implicit bias in function space. The authors demonstrate the solution of training shallow ReLU networks with different widths and show that the solution function closely fits the training data and has the smallest 2-norm of the second derivative weighted by a curvature penalty. They also analyze the impact of various initialization procedures on the resulting functions. The findings can be applied to both univariate and multivariate regression tasks.
JOURNAL OF MACHINE LEARNING RESEARCH
(2023)
Article
Computer Science, Interdisciplinary Applications
M. Taavoni, M. Arashi, Wan-Lun Wang, Tsung- Lin
Summary: This paper introduces a new multivariate semiparametric mixed model (MtSMM) which combines a parametric linear function, random effects, and a nonparametric smooth function to improve robustness against outliers, providing greater flexibility in analyzing longitudinal trajectories. Simulation studies and a real example on PBCseq data demonstrate the empirical behavior of the proposed methodology.
JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION
(2021)
Article
Biology
Jiakun Jiang, Wei Yang, Erin M. Schnellinger, Stephen E. Kimmel, Wensheng Guo
Summary: In this article, a dynamic logistic state space model is proposed to continuously update parameters for better prediction accuracy. The model allows for both time-varying and time-invariant coefficients, with time-varying coefficients modeled using smoothing splines and smoothing parameters chosen by maximum likelihood. The model is updated using batch data to approximate the underlying binomial density function. Simulation results show significantly higher prediction accuracy compared to existing methods. The method is applied to predict 1 year survival after lung transplantation using United Network for Organ Sharing data.
Article
Statistics & Probability
Javier Albert-Smet, Aurora Torrente, Juan Romo
Summary: The k-Means algorithm is a popular choice for clustering data, but it is known to be sensitive to the initialization process. This paper introduces an extension to the BRIk algorithm for longitudinal data, which clusters centroids derived from bootstrap replicates of the data and utilizes the Modified Band Depth. The proposed approach enhances the BRIk method by fitting B-splines to observations and incorporating a resampling process, resulting in improved effectiveness in providing initial seeds for k-Means clustering.
ADVANCES IN DATA ANALYSIS AND CLASSIFICATION
(2023)
Article
Chemistry, Analytical
Sufyan Ali Memon, Min-Seuk Park, Imran Memon, Wan-Gu Kim, Sajid Khan, Yifang Shi
Summary: This research extends the FIsJITS filter to address the unknown dynamics of multi-maneuvering targets (MMT) in tracking scenarios, and proposes the MMT-sJITS method. By optimizing the estimation of target measurements concealed by joint measurements, this method reduces computational complexity and improves tracking performance.
Article
Meteorology & Atmospheric Sciences
Reinhold Steinacker
Summary: This article introduces a new approach of splines to examine climatological time series and demonstrates various applications. Instead of the classical spline-procedure, a variational formulation is used which minimizes a cost function to obtain smooth interpolated values at arbitrary points. The method innovatively selects the constraint of conserving mean values within specific smoothing intervals. It enables accurate derivation of past and present climate conditions, refined computation of extreme values and probability of temperature thresholds, detection of break points in climatological time series, and consistent interpolation of mean-only time series. The article presents and discusses several applications of the method using climatological time series of Vienna.
INTERNATIONAL JOURNAL OF CLIMATOLOGY
(2023)
Article
Chemistry, Analytical
Wen-Hao Zhang, Lin Dai, Wang Chen, Anyu Sun, Wu-Le Zhu, Bing-Feng Ju
Summary: In this paper, a novel information-driven smoothing spline linearization method is proposed for high-precision displacement sensors. The proposed method outperforms traditional methods and achieves better linearization results.
Article
Geosciences, Multidisciplinary
Peiyu Miao, Genru Xiao, Shengping Wang, Keliang Zhang, Buang Bai, Zeng Guo
Summary: This study examined the effects of different seasonal fitting techniques on the spatial distribution of common mode errors using the coordinate time series of GPS reference stations in China. The results showed that smoothing spline fitting outperformed constant amplitude harmonic fitting and continuous wavelet transform in fitting short-term irregular seasonal signals. Additionally, the residual time series obtained from smoothing spline fitting had the lowest root mean square error and exhibited the best spatial filtering effect.
FRONTIERS IN EARTH SCIENCE
(2023)
Article
Automation & Control Systems
Jiajin Wei, Chen Zhu, Zhi-Min Zhang, Ping He
Summary: This paper reviews several iteratively reweighted baseline correction methods and proposes a new method called two-stage iteratively reweighted smoothing splines (RWSS) to address the estimation of baselines in complex-structured signals. Simulation studies and real data experiments demonstrate the performance and reliability of the proposed method.
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS
(2022)
Article
Computer Science, Software Engineering
Yuriy Korablev
Summary: This paper describes a cubic integral smoothing spline method for restoring a function using integrals, and provides a detailed mathematical method for constructing such a spline. The method is based on a cubic integral spline with a penalty function, allowing control over the smoothness of the restored function and the nonlinearity of the spline. An implementation in the R language is given, along with example applications.
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE
(2022)
Article
Statistics & Probability
Chin-Shang Li, Minggen Lu
Summary: A spline-based zero-inflated Bernoulli (ZIB) regression model is proposed to capture potentially nonlinear effects of continuous covariates. The study demonstrates that under a smoothness condition, the spline estimator is consistent and the regression parameter estimators are asymptotically normally distributed.
JOURNAL OF APPLIED STATISTICS
(2022)
