A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z  

  Subjects -> STATISTICS (Total: 130 journals)
The end of the list has been reached or no journals were found for your choice.
Similar Journals
Journal Cover
Journal of the Korean Statistical Society
Journal Prestige (SJR): 0.545
Citation Impact (citeScore): 1
Number of Followers: 0  
 
  Hybrid Journal Hybrid journal (It can contain Open Access articles)
ISSN (Print) 1226-3192 - ISSN (Online) 2005-2863
Published by Elsevier Homepage  [2971 journals]
  • A Bayesian method for multinomial probit model

    • Free pre-print version: Loading...

      Abstract: Abstract The independence of irrelevant alternatives (IIA) property states that the ratio of any two choice probabilities in a set of alternatives is independent of the presence or absence of other alternatives. In the modeling of multinomial data, the IIA is not feasible. In this situation, the multinomial probit (MNP) model is a type of discrete choice model that is commonly used. Due to the identifiability problem and the positive-definiteness constraint, modeling the covariance matrix in the MNP is difficult. All existing methods use unidentifiable parameters in the covariance matrix to solve the unidentifiability problem and improve the rate of convergence of a data augmentation algorithm. These methods also use the inverse Wishart distribution, which is frequently insufficient (Barnard et al. Stat Sin 10(4):1281–1311, 2000). We employed variance-correlation decomposition to decompose the identifiable covariance matrix into standard deviations and a correlation matrix instead of using the unidentifiable covariance matrix. Hypersphere decomposition was also used to decompose the correlation matrix. Thus, the estimated covariance matrix satisfied the positive definiteness constraint. The performance of our proposed model was illustrated using a detergent dataset from market research.
      PubDate: 2022-12-03
       
  • A note on maximum likelihood estimation for mixture models

    • Free pre-print version: Loading...

      Abstract: Abstract Practitioners as well as some statistics students often blindly use standard software or algorithms to get maximum likelihood estimator (MLE) without checking the validity of existence of such an estimator. Even in simple situations where data comes from mixtures of Gaussians, global MLE does not exist. This note is intended as a teachers corner, highlighting existential issues related to MLE for mixture models, even when the components are not necessarily Gaussian.
      PubDate: 2022-12-01
       
  • Estimation of the parameters of a Wishart extension on symmetric matrices

    • Free pre-print version: Loading...

      Abstract: Abstract This paper deals with the parameters of a natural extension of the Wishart distribution, that is the Riesz distribution on the space of symmetric matrices. We estimate the shape parameter using two different approaches. The first one is based on the method of moments, we give its expression and investigate some of its properties. The second represents the maximum likelihood estimator. Unfortunately, in this case we do not have an explicit formula for this estimator. This latter is expressed in terms of the digamma function and sample mean of log-gamma variables. However, we derive the strong consistency and asymptotic normality properties of this estimator. A numerical comparative study between the two estimators is carried out in order to test the performance of the proposed approaches. For the second parameter, that is the scale parameter, we prove that the distribution of the maximum likelihood estimator given by Kammoun et al. (J Statist Prob Lett 126:127–131, 2017) is related to the Riesz distribution. We examine some properties concerning this estimator and we assess its performance by a numerical study.
      PubDate: 2022-12-01
       
  • A computationally efficient and flexible algorithm for high dimensional
           mean and covariance matrix change point models

    • Free pre-print version: Loading...

      Abstract: Abstract This paper proposes a computationally efficient algorithm, FBS (Fast Binary Segmentation), for both single and multiple change point detection under high-dimensional setups. As a general technique, it can be widely used in various change point problems including mean vectors and covariance matrices change point models. Based on various \(\ell _{(s,p)}\) -norm aggregations for the cumulative sum (CUSUM) statistics, the new algorithm can be applied to a wide range of alternative structures including sparse and dense settings as special cases. We present the essence of the new algorithm. The efficiency and accuracy of the new algorithm are justified via comparing it with the other existing techniques. Lastly, a real data application further demonstrates the usefulness of our method.
      PubDate: 2022-12-01
       
  • Bayesian empirical likelihood inference for the generalized binomial AR(1)
           model

    • Free pre-print version: Loading...

      Abstract: Abstract In this paper, the Bayesian empirical likelihood (BEL) inference is considered for the generalized binomial AR(1) model. We establish a nonparametric likelihood using the empirical likelihood (EL) approach and consider a specific prior based on copulas. An efficient Markov chain Monte Carlo (MCMC) procedure is described for the required computation of the posterior distribution. In the simulation study, we analyze the accuracy and sensitivity of the MCMC algorithm. We also study the robustness of the new method. The results imply that our algorithm converges quickly and not strongly influenced by the model assumptions. Furthermore, the BEL method is robust. Finally, a real data example is analyzed to illustrate of our method.
      PubDate: 2022-12-01
       
  • Flexible INAR(1) models for equidispersed, underdispersed or overdispersed
           counts

    • Free pre-print version: Loading...

      Abstract: Abstract Equidispersed, underdispersed and overdispersed count data are commonly encountered in practice. To better describe these data characteristics, this paper develops two classes of INAR(1) processes, which not only can model a wide range of overdispersion and underdispersion, but also have ability to describe the zero-inflated and zero-deflated characteristics of the count data. The probabilistic and statistical properties of the two processes are studied. Estimators of the model parameters are derived by using conditional maximum likelihood (CML) and modified conditional least squares (MCLS) methods. Some asymptotic properties and numerical results of the estimators are investigated. Three real examples are given to show the flexibility and usefulness of the proposed models.
      PubDate: 2022-12-01
       
  • Robust estimation for a general functional single index model via quantile
           regression

    • Free pre-print version: Loading...

      Abstract: Abstract This paper studies the estimation of a general functional single index model, in which the conditional distribution of the response depends on the functional predictor via a functional single index structure. We find that the slope function can be estimated consistently by the estimation obtained by fitting a misspecified functional linear quantile regression model under some mild conditions. We first obtain a consistent estimator of the slope function using functional linear quantile regression based on functional principal component analysis, and then employ a local linear regression technique to estimate the conditional quantile function and establish the asymptotic normality of the resulting estimator for it. The finite sample performance of the proposed estimation method is studied in Monte Carlo simulations, and is illustrated by an application to a real dataset.
      PubDate: 2022-12-01
       
  • Revisiting feature selection for linear models with FDR and power
           guarantees

    • Free pre-print version: Loading...

      Abstract: Abstract The problem of feature selection for linear models is re-examined by using the fixed-X knockoff procedure and incorporating the selection probability as variable importance scores. Unlike previous work that predominantly focused on false discovery rate (FDR) control, this paper aims to establish theoretical power guarantees for the fixed-X knockoff procedure in linear models. An intersection selection procedure is proposed to make better use of sample data for estimating the selection probability, which in practice results in increasing the selection power. In addition, a two-stage procedure by using the data splitting technique is further developed to explore related theoretical results under high dimensionality. The performance of the proposal over its main competitors is demonstrated through comprehensive simulation studies and real data analysis.
      PubDate: 2022-12-01
       
  • Penalized polygram regression

    • Free pre-print version: Loading...

      Abstract: Abstract We consider a study on regression function estimation over a bounded domain of arbitrary shapes based on triangulation and penalization techniques. A total variation type penalty is imposed to encourage fusion of adjacent triangles, which leads to a partition of the domain consisting of disjointed polygons. The proposed method provides a piecewise linear, and continuous estimator over a data adaptive polygonal partition of the domain. We adopt a coordinate decent algorithm to handle the non-separable structure of the penalty and investigate its convergence property. Regarding the asymptotic results, we establish an oracle type inequality and convergence rate of the proposed estimator. A numerical study is carried out to illustrate the performance of this method. An R software package polygram is available.
      PubDate: 2022-12-01
       
  • Partial linear regression of compositional data

    • Free pre-print version: Loading...

      Abstract: Abstract We study a partial linear model in which the response is compositional and the predictors include both compositional and Euclidean variables. We define a partial linear regression model under Aitchison geometry based on isometric log-ratio (ilr) transformation. An identification condition of linear parameters is provided in terms of expectations and conditional expectations of the response and covariates. An estimator based on the identification is developed and asymptotic properties of proposed estimators are derived. The proposed method can be implemented easily by using existing R packages such as np and compositions. The limiting distribution of the proposed estimator is provided with normal distribution in Euclidean space so that it is easy to use for inference. Also, some finite sample properties are presented via simulation studies. We also present election data as an illustrative example.
      PubDate: 2022-12-01
       
  • Robust coefficients of correlation or spatial autocorrelation based on
           implicit weighting

    • Free pre-print version: Loading...

      Abstract: Abstract Pearson product-moment correlation coefficient represents a fundamental tool for measuring linear association between two data vectors. In various applications, it is often reasonable to consider its weighted version known as the weighted correlation coefficient. This paper starts with theoretical considerations related to properties of the weighted correlation coefficient, particularly to its local robustness and relationship to other similarity measures. Inspired by the least weighted squares regression estimator, a robust correlation coefficient is investigated here together with its spatial autocorrelation extension. Finally, the considered methods are investigated in two image processing tasks.
      PubDate: 2022-12-01
       
  • Wild bootstrap Ljung–Box test for residuals of ARMA models robust to
           variance change

    • Free pre-print version: Loading...

      Abstract: Abstract Ljung–Box (LB) test is one of the most popular test for determining whether autocorrelations in residuals of fitted time series models exist or not. However, it may not be appropriate to apply LB test to time series data with variance change due to size distortions. In this paper, we proposed a wild bootstrap-based LB test for residuals of fitted ARMA models. Our simulation study shows that our wild bootstrap-based LB test achieves the correct sizes and comparable powers in finite samples in the presence of variance change.
      PubDate: 2022-12-01
       
  • Monitoring multivariate data with high missing rate by pooling univariate
           statistics

    • Free pre-print version: Loading...

      Abstract: Abstract In this paper, we propose a control chart to monitor multivariate data with missing values as an alternative to the most common imputation-based control chart. The chart statistic we use is the weighted sum of the chi-square statistics of each variable, say \(Q^*\) , which is named as pooled component test statistic (PCT) by Wu et al. (2006). We modify the statistic in the context of monitoring and approximate the in-control distribution of \(Q^*\) as a scaled chi-square distribution. The PCT chart we propose in this paper is a Shewhart chart using \(Q^*\) , and its control limits are from the estimate of the approximate in-control distribution of \(Q^*\) . We numerically show that the PCT chart performs better than the imputation-based methods in the literature. We finally apply it to monitoring a semiconductor manufacturing process using production environmental variables.
      PubDate: 2022-12-01
       
  • Projective resampling estimation of informative predictor subspace for
           multivariate regression

    • Free pre-print version: Loading...

      Abstract: Abstract In this paper, a paradigm to estimate the so-called informative predictor subspace (Yoo in Statistics 50:1086–1099, 2016) for multivariate regression is proposed. For this, as a primary target subspace, a projective resampling informative predictor subspace is newly developed. The projective resampling informative predictor subspace is constructed based on a projection resampling method by Li et al. (2008), and it has advantage that it is smaller than the original informative predictor subspace but contains the central subspace. To estimate the new target subspace, the three approaches of projective resampling, coordinate, and coordinate-projective resampling mean methods are proposed. The three methods are investigated via various numerical studies, which confirm their potential usefulness in practice.
      PubDate: 2022-12-01
       
  • Autocovariance estimation in the presence of changepoints

    • Free pre-print version: Loading...

      Abstract: Abstract This article studies estimation of a stationary autocovariance structure in the presence of an unknown number of mean shifts. Here, a Yule–Walker moment estimator for the autoregressive parameters in a dependent time series contaminated by mean shift changepoints is proposed and studied. The estimator is based on first order differences of the series and is proven consistent and asymptotically normal when the number of changepoints m and the series length N satisfy \(m/N \rightarrow 0\) as \(N \rightarrow \infty\) .
      PubDate: 2022-12-01
       
  • Non-parametric comparison and classification of two large-scale
           populations

    • Free pre-print version: Loading...

      Abstract: Abstract In this paper, we investigate a non-parametric approach to compare two groups in microarray data. This is done using a threshold penalized-distance likelihood function, which is made up of a penalty and a suitable threshold distance, and is applicable when sample size is small or when the data is not normally distributed. We also use this function to classify new data. This is based on objects that are identified as differences between the two groups, not for all objects. We also study a real data application to illustrate our methods.
      PubDate: 2022-11-21
       
  • Estimating mixed-memberships using the symmetric laplacian inverse matrix

    • Free pre-print version: Loading...

      Abstract: Abstract Mixed membership community detection is a challenging problem. In this paper, to detect mixed memberships, we propose a new method Mixed-SLIM which is a spectral clustering method on the symmetrized Laplacian inverse matrix under the degree-corrected mixed membership model. We provide theoretical bounds for the estimation error on the proposed algorithm and its regularized version under mild conditions. Meanwhile, we provide some extensions of the proposed method to deal with large networks in practice. These Mixed-SLIM methods outperform state-of-art methods in simulations and substantial empirical datasets for both community detection and mixed membership community detection problems.
      PubDate: 2022-11-21
       
  • The modified Yule-Walker method for multidimensional infinite-variance
           periodic autoregressive model of order 1

    • Free pre-print version: Loading...

      Abstract: Abstract The time series with periodic behavior, such as the periodic autoregressive (PAR) models belonging to the class of the periodically correlated processes, are present in various real applications. In the literature, such processes were considered in different directions, especially with the Gaussian-distributed noise. However, in most of the applications, the assumption of the finite-variance distribution seems to be too simplified. Thus, one can consider the extensions of the classical PAR model where the non-Gaussian distribution is applied. In particular, the Gaussian distribution can be replaced by the infinite-variance distribution, e.g. by the \(\alpha\) -stable distribution. In this paper, we focus on the multidimensional \(\alpha\) -stable PAR time series models. For such models, we propose a new estimation method based on the Yule-Walker equations. However, since for the infinite-variance case the covariance does not exist, thus it is replaced by another measure, namely the covariation. In this paper we propose to apply two estimators of the covariation measure. The first one is based on moment representation (moment-based) while the second one—on the spectral measure representation (spectral-based). The validity of the new approaches are verified using the Monte Carlo simulations in different contexts, including the sample size and the index of stability of the noise. Moreover, we compare the moment-based covariation-based method with spectral-based covariation-based technique. Finally, the real data analysis is presented.
      PubDate: 2022-10-21
       
  • Correction: Robust MAVE for single-index varying-coefficient models

    • Free pre-print version: Loading...

      PubDate: 2022-09-16
       
  • Robust MAVE for single-index varying-coefficient models

    • Free pre-print version: Loading...

      Abstract: Abstract In this paper, a robust, efficient and easily implemented estimation procedure for single-index varying-coefficient models is proposed by combining minimum average variance estimation (MAVE) with exponential squared loss. The merit of the proposed method is robust against outliers or heavy-tailed error distributions while asymptotically efficient as the original MAVE under the normal error case. A practical minorization–maximization algorithm is proposed for implementation. Under some regularity conditions, asymptotic distributions of the resulting estimators are derived. Simulation studies and a real data example are conducted to examine the finite sample performance of the proposed method. Both theoretical and empirical findings confirm that our proposed method works very well.
      PubDate: 2022-09-02
       
 
JournalTOCs
School of Mathematical and Computer Sciences
Heriot-Watt University
Edinburgh, EH14 4AS, UK
Email: journaltocs@hw.ac.uk
Tel: +00 44 (0)131 4513762
 


Your IP address: 44.210.85.190
 
Home (Search)
API
About JournalTOCs
News (blog, publications)
JournalTOCs on Twitter   JournalTOCs on Facebook

JournalTOCs © 2009-