Article
Computer Science, Artificial Intelligence
Jiaming Zhang, Hanwen Ning, Xingjian Jing, Tianhai Tian
Summary: A novel approach for addressing the issue of fixed bandwidth limitation in online learning is proposed, which achieves adaptive learning and parameter updates through a linearization scheme and optimal control techniques. Compared to fixed bandwidth methods, the new approach provides better prediction accuracy and faster convergence speed.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Management
Xiaoyu Liu, Xing Yan, Kun Zhang
Summary: This study focuses on the use of a kernel quantile estimator (KQE) for nested simulation to estimate value at risk (VaR) in portfolio risk measurement. The bias, variance, and mean squared error (MSE) are analyzed, and an efficient bootstrap-based algorithm is proposed for practical implementation.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2024)
Article
Statistics & Probability
Fabienne Comte, Nicolas Marie
Summary: In this study, a bandwidth selection method is proposed for estimating the regression function using kernel estimators. The simulation study demonstrated that the single-bandwidth cross-validation estimator performs better in small noise context.
ELECTRONIC JOURNAL OF STATISTICS
(2021)
Article
Computer Science, Interdisciplinary Applications
Nasim Abdolmaleki, Leyli Mohammad Khanli, Mahdi Hashemzadeh, Shahin Pourbahrami
Summary: This paper introduces an Apollonius Circle-based Quantum Clustering (ACQC) method, which adaptively sets the kernel bandwidth without prior knowledge, resulting in improved accuracy and efficiency in clustering.
JOURNAL OF COMPUTATIONAL SCIENCE
(2022)
Article
Computer Science, Artificial Intelligence
Cuiling Chen, Jian Wei, Zhi Li
Summary: The purpose of multiple kernel clustering (MKC) is to generate an optimal kernel by incorporating information from multiple base kernels. However, previous methods either use a neighborhood kernel to enlarge the search range or select local base kernels to avoid redundancy, without combining both methods. In this paper, a new method called ONKC-ALK-BD is proposed to overcome these limitations. It uses a weight strategy to select local base kernels and applies a block diagonal regularizer to encourage a block diagonal structure in the clustering indicator matrix. Experimental results on twelve datasets demonstrate the effectiveness of the proposed method.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Computer Science, Interdisciplinary Applications
Pablo Martinez-Camblor
Summary: Binary classification problems play a crucial role in biomedicine, involving diagnostic and personalized medicine. We propose an estimation procedure for optimal classification rules and demonstrate its practical use in a real-world problem.
JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION
(2022)
Article
Engineering, Mechanical
R. Cumbo, L. Mazzanti, T. Tamarozzi, P. Jiranek, W. Desmet, F. Naets
Summary: The paper presents a Kalman-based methodology for inverse load identification, introducing two alternative metrics to improve sensor selection and enhance the accuracy of estimated quantities.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Statistics & Probability
Taha Hussein Ali
Summary: This research proposes a new improvement for the Nadaraya-Watson kernel nonparametric regression estimator. The bandwidth of this improvement is obtained based on statistical indicators such as robust mean, median, and harmonic mean of kernel function. Simulation study shows that the proposed estimator, especially when using harmonic mean, outperforms classical methods in terms of accuracy.
COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION
(2022)
Article
Mathematics
Necla Gunduz, Sule Karakoc
Summary: This research conducts a comparative analysis of seven bandwidth calculation techniques, finding that the BCV method provides the closest result to the optimal bandwidth for normal distributions, while the SCV method is suited for real-world wind speed data.
Article
Mathematical & Computational Biology
Wende Clarence Safari, Ignacio Lopez-de-Ullibarri, Maria Amalia Jacome
Summary: In this study, a nonparametric estimator for the conditional survival function in the mixture cure model is introduced for right-censored data with partial knowledge of cure status. The estimator, developed for a single continuous covariate, can be extended to multiple covariates. Results show a reduction in variance compared to previous estimators and better performance across a range of covariate values when the bandwidth parameter is suitably chosen.
BIOMETRICAL JOURNAL
(2021)
Article
Statistics & Probability
Necla Gunduz, Celal Aydin
Summary: This study provides simulation-based exploration and characterization of key kernel density functionals, introduces Cauchy-scale estimators as an alternative approach for preliminary bandwidth estimation, and analyzes the sampling distribution of bandwidth estimators. The proposed method shows lower variance in both simulations and real data applications.
JOURNAL OF APPLIED STATISTICS
(2021)
Article
Computer Science, Information Systems
Shengyu Zhu, Biao Chen, Zhitang Chen, Pengfei Yang
Summary: This study characterizes the asymptotic performance of nonparametric one- and two-sample testing, focusing on optimal tests and error exponents. Sanov's theorem is used to derive sufficient conditions for achieving optimal error exponents in one-sample tests, while MMD and KSD based tests are shown to achieve optimal performance in specific conditions. An extended version of Sanov's theorem is established for general two-sample testing to derive tests with optimal error exponents under given constraints.
IEEE TRANSACTIONS ON INFORMATION THEORY
(2021)
Article
Computer Science, Artificial Intelligence
Jian Zou, Yue Zhang, Hongjian Liu, Lifeng Ma
Summary: This paper presents a novel method for single sample face recognition using grayscale monogenic features and kernel sparse representation on multiple Riemannian manifolds. The approach involves extracting local features from different regions of the face images, modeling the corresponding feature vectors as points on a Grassmann manifold, extracting co-occurrence distributions of feature images, and training a kernel sparse representation classifier using multiple kernel fusion. Experimental results demonstrate the superiority of the proposed method.
Article
Thermodynamics
Chang-uk Ahn, Chanhun Park, Dong Il Park, Jin-Gyun Kim
Summary: This paper proposes an effective iterative hybrid parameter selection algorithm to obtain stable inverse solutions by combining the regularization parameter alpha and the hybrid parameter beta to control the amplification error of the inverse algorithm. The initial alpha is defined by computing the sum of the bias and variance errors, and the total error can be reduced by adjusting beta to achieve better stability of the inverse solutions.
INTERNATIONAL JOURNAL OF HEAT AND MASS TRANSFER
(2022)
Article
Computer Science, Information Systems
Yu-Lin He, Xuan Ye, De-Fa Huang, Joshua Zhexue Huang, Jun-Hai Zhai
Summary: This paper proposes a novel ensemble UCV based KDE (EUCV-KDE), which determines the expectation of an estimated PDF using an ensemble of data-block based UCVs. A novel objective function is designed for EUCV-KDE by considering the empirical and structural risk of KDE together. The experimental results show that EUCV-KDE is more stable and performs better than classical UCV-KDE and RCV-KDE.
INFORMATION SCIENCES
(2021)
Article
Statistics & Probability
Rebeca Pelaez Suarez, Ricardo Cao Abad, Juan M. Vilar Fernandez
Summary: This paper proposes and compares four nonparametric estimators of default probability in credit risk, derived from estimators of the conditional survival function for censored data. The performance of these estimators is demonstrated through simulation and empirical studies.
Article
Computer Science, Cybernetics
Jorge Perez-Martin, Alejandro Rodriguez-Ascaso, Elisa M. Molanes-Lopez
Summary: Students using YouTube's automatic speech recognition feature to produce captions made errors mainly in the number of characters per line, caption speed, failure to use a new line per participant, and not including sound effects. Recommendations for improvement include providing more guidance in the course, teaching how to edit captions and add sound effects subtitles.
UNIVERSAL ACCESS IN THE INFORMATION SOCIETY
(2021)
Correction
Statistics & Probability
Rebeca Pelaez Suarez, Ricardo Cao Abad, Juan M. Vilar Fernandez
Article
Statistics & Probability
Laura Borrajo, Ricardo Cao
Summary: This paper investigates nonparametric estimation for a large-sized sample subject to sampling bias, proposing a new method that integrates kernel density estimation and outperforms classical methods in mean estimation. Simulation results show the positive performance of the new method with suitable choices of smoothing parameters, as well as the influence of these parameters on the final estimator.
Article
Orthopedics
Ricardo Larrainzar-Garijo, Elisa M. Molanes-Lopez, David Murillo-Vizuete, Raul Garcia-Bogalo, David Escobar-Anton, Jesus Lopez-Rodriguez, Angel Diez-Fernandez, Fernando Corella-Montoya
Summary: This study investigated the dynamic coronal HKA angle after mechanically aligned knee replacement, finding significant differences between preoperative and postoperative statuses, with 98.6% of knees being within +/- 3 degrees of the HKA at full extension and 80.6% achieving within-range postoperative dynamic alignment at any grade of flexion.
JOURNAL OF KNEE SURGERY
(2022)
Article
Immunology
Christian Vaquero-Yuste, Ignacio Juarez, Marta Molina-Alejandre, Elisa Maria Molanes-Lopez, Adrian Lopez-Nares, Fabio Suarez-Trujillo, Alberto Gutierrez-Calvo, Adela Lopez-Garcia, Inmaculada Lasa, Remedios Gomez, Eduardo Fernandez-Cruz, Carmen Rodrigez-Sainz, Antonio Arnaiz-Villena, Jose Manuel Martin-Villa
Summary: A study on 107 Spanish patients with gastric adenocarcinoma and 58 healthy controls revealed that the frequency of the 14bp DEL allele in the HLA-G gene was significantly increased in patients, as well as the haplotype formed by the combination of 14bp DEL/+3142 C variants. Patients with the 14bp DEL/DEL genotype showed a lower 5-year life expectancy compared to other genotypes. The findings suggest a potential association between HLA-G gene polymorphisms and gastric adenocarcinoma susceptibility, disease progression, and survival.
FRONTIERS IN IMMUNOLOGY
(2021)
Article
Mathematics
Rebeca Pelaez, Ricardo Cao, Juan M. Vilar
Summary: This paper proposes an algorithm based on resampling methods to estimate the probability of default and calculate the confidence intervals. The method shows good behavior in extensive simulation studies and is applied to analyze a German credit dataset.
Article
Orthopedics
Ricardo Larrainzar-Garijo, Elisa M. Molanes-Lopez, Miguel Canones-Martin, David Murillo-Vizuete, Natalia Valencia-Santos, Raul Garcia-Bogalo, Fernando Corella-Montoya
Summary: This study aimed to investigate whether the use of a surgical navigation system in total knee replacement (TKR) can enable beginner and intermediate surgeons to achieve clinical outcomes comparable to those achieved by expert surgeons in the long term. The results showed that under expert guidance, beginner and intermediate surgeons could achieve comparable accuracy of implant positioning, limb alignment, and long-term clinical outcomes in navigated assisted TKR.
INDIAN JOURNAL OF ORTHOPAEDICS
(2022)
Article
Cell Biology
Ana Aguinaga-Barrilero, Ignacio Juarez, Christian Vaquero-Yuste, Marta Molina-Alejandre, Alberto Gutierrez-Calvo, Inmaculada Lasa, Adela Lopez, Remedios Gomez, Elisa M. Molanes-Lopez, Jose M. Martin-Villa
Summary: The presence of cells with regulatory functions in patients with cancer is one of the mechanisms whereby the immune system cannot confront tumor growth. In patients with gastric adenocarcinoma, CD8+LAP+ cells are increased in tumoral sites, especially in early stages of the disease.
CELLULAR IMMUNOLOGY
(2022)
Editorial Material
Statistics & Probability
Ricardo Cao
Summary: This paper discusses the invited paper by Lopez-Cheda, Peng and Jacome on nonparametric mixture cure models with covariates. An alternative estimation procedure is proposed. The situation when the two covariate vectors share some but not all their covariates is also considered. Technical aspects in the assumptions, results and proofs of the invited paper are also discussed. Comments on the simulations and the real-data application are included. Possible interesting topics for further research in this field are briefly discussed.
Article
Public, Environmental & Occupational Health
Ines Barbeito, Daniel Precioso, Maria Jose Sierra, Susana Vegas-Azcarate, Sonia Fernandez Balbuena, Begona Vitoriano, David Gomez-Ullate, Ricardo Cao, Susana Monge
Summary: This study estimated the association between the level of restriction in different fields of activity and SARS-CoV-2 transmission in Spain. The results showed that increasing restrictions can reduce COVID-19 transmission, especially in the areas of culture and leisure, social distancing, indoor restaurants, and indoor sports. However, there are limitations in the study, including collinearity between fields and the artificial quantification of qualitative restrictions, so caution is needed in attributing the effects to specific areas.
FRONTIERS IN PUBLIC HEALTH
(2023)
Article
Multidisciplinary Sciences
Beatriz Pineiro-Lamas, Ana Lopez-Cheda, Ricardo Cao, Laura Ramos-Alonso, Gabriel Gonzalez-Barbeito, Cayetana Barbeito-Caamano, Alberto Bouzas-Mosquera
Summary: This dataset is a result of collaboration between two universities, containing information about 531 women with HER2+ breast cancer treated with potentially cardiotoxic therapies. The dataset aims to enable research on cardiovascular side effects and early detection of cardiac problems.
Article
Environmental Sciences
Noelia Trigo-Tasende, Juan A. Vallejo, Soraya Rumbo-Feal, Kelly Conde-Perez, Manuel Vaamonde, Angel Lopez-Oriona, Ines Barbeito, Mohammed Nasser-Ali, Ruben Reif, Bruno K. Rodino-Janeiro, Elisa Fernandez-Alvarez, Iago Iglesias-Corras, Borja Freire, Javier Tarrio-Saavedra, Laura Tomas, Pilar Gallego-Garcia, David Posada, German Bou, Ignacio Lopez-de-Ullibarri, Ricardo Cao, Susana Ladra, Margarita Poza
Summary: The wastewater-based epidemiology program COVIDBENS in A Coruna, Spain, monitored viral load and detected SARS-CoV-2 mutations in wastewater using RT-qPCR and Illumina sequencing. It successfully predicted community outbreaks and identified new variants, providing early warning to local authorities and health managers.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Computer Science, Information Systems
Elisa M. Molanes-Lopez, Alejandro Rodriguez-Ascaso, Emilio Leton, Jorge Perez-Martin
Summary: The study addressed the challenge of multimedia accessibility assessment by involving a group of novice evaluators and applying a set of criteria for assessment.
Article
Public, Environmental & Occupational Health
Ana Lopez-Cheda, Maria-Amalia Jacome, Ricardo Cao, Pablo M. De Salazar
Summary: This study focuses on modeling the lengths-of-stay of hospitalized COVID-19 patients using real-time surveillance data, demonstrating that a non-parametric mixture cure model outperforms standard methods in estimating ICU and HW lengths-of-stay, and emphasizing the importance of adjusting for sex and age in accurately predicting occupancy rates and discharge/death outcomes.
EPIDEMIOLOGY AND INFECTION
(2021)
