When reporting the results of clinical studies, some researchers may choose five‐number summary (including sample median, first and third quartiles, minimum maximum values) rather than mean standard deviation (SD), particularly for skewed data. For these when included in a meta‐analysis, it is often desired to convert back SD. this purpose, several methods have been proposed recent literature they are increasingly used nowadays. In article, we propose further advance by developing smoothly...

A Hawkes process is also known under the name of a self-exciting point and has numerous applications throughout science engineering. We derive statistical estimation (maximum likelihood estimation) goodness-of-fit (mainly graphical) for multivariate processes with possibly dependent marks. As an application, we analyze two data sets from finance.

A Hawkes process is also known under the name of a self-exciting point and has numerous applications throughout science engineering. We derive statistical estimation (maximum likelihood estimation) goodness-of-fit (mainly graphical) for multivariate processes with possibly dependent marks. As an application, we analyze two data sets from finance.

In microbiome studies, an important goal is to detect differential abundance of microbes across clinical conditions and treatment options. However, the compositional data (quantified by relative abundance) are highly skewed, bounded in [0, 1), often have many zeros. A two-part model commonly used separate zeros positive values explicitly two submodels: a logistic for probability specie being present Part I, Beta regression conditional on presence II. coefficients II cannot provide marginal...

Carbon emissions based on land use change have attracted extensive attention from scholars, but the current carbon emission accounting model is still relatively rough. Despite continuous promotion of China’s ecological civilization strategy, whether green economic development promotes reduction remains to be studied. This study uses Exploratory Spatial-temporal Data Analysis (ESTDA) framework system revise land-use model; it integrates NDVI adjustment index and systematically analyzes...

Motivated by increment process modeling for two correlated random and non-random systems from a discrete-time asset pricing with both risk free risky security, we propose class of semiparametric regressions combination system. Unlike classical regressions, mean regression functions in the new model contain variance components variables are related to latent variables, which certain economic interpretation can be made. The motivating example explains why GARCH-M function contains component...

This article investigates the possible use of our newly defined extended projection depth (abbreviated to EPD) in nonparametric discriminant analysis. We propose a robust classifier, which relies on intuitively simple notion EPD. The EPD-based classifier assigns an observation population with respect it has maximum Asymptotic properties misclassification rates and are discussed. A few simulated data sets used compare performance Fisher's linear rule, quadratic PD-based classifier. It is also...

In the era of big data, one important issues is how to recover sets true features when data arrive sequentially. The paper presents a general framework for online updating variable selection and parameter estimation in generalized linear models with streaming datasets. This type penalized likelihoods differentiable or non-differentiable penalty functions. An coordinate descent algorithm proposed solving optimization problem. Moreover, tuning suggested an way. consistencies oracle property...

This paper develops a robust estimation procedure for the varying-coefficient partially linear model via local rank technique. The new provides highly efficient and alternative to least-squares method. In other words, proposed method is across wide class of non-normal error distributions it only loses small amount efficiency normal error. Moreover, test hypothesis constancy nonparametric component proposed. statistic simple thus can be easily implemented. We conduct Monte Carlo simulation...

We develop a consistent and highly efficient marginal model for missing at random data using an estimating function approach. Our approach differs from inverse weighted equations (Robins, Rotnitzky, Zhao 1995) the imputation method (Paik 1997) in that our does not require probability of or imputing response based on assumed models. The proposed is aggregate unbiased approach, which likelihood function; however, it equivalent to score equation if known. aggregate-unbiased best linear...

We propose a penalized quantile regression for partially linear varying coefficient (VC) model with longitudinal data to select relevant non parametric and components simultaneously. Selection consistency oracle property are established. Furthermore, if part VC unknown, we new unified method, which can do three types of selections: separation constant effects, selection variables, it be carried out conveniently in one step. Consistency the selections estimation established as well....

Markov chain Monte Carlo (MCMC) methods can be used for statistical inference. The are time-consuming due to time-vary. To resolve these problems, parallel tempering (PT), as a MCMC method, is tried, dynamic generalized linear models (DGLMs), well the several optimal properties of our proposed method. In PT, two or more samples drawn at same time, and exchange information with each other. We also present some simulations DGLMs in case provide applications Poisson-type financial research.

Differenced estimators of variance bypass the estimation regression function and thus are simple to calculate. However, there exist two problems: most differenced do not achieve asymptotic optimal rate for mean square error; finite samples bias is also important further considered. In this paper, we estimate as intercept in a linear with lagged Gasser-type estimator dependent variable. For equidistant design, our only $$n^{1/2}$$ -consistent asymptotically normal, but achieves bound terms...

Partial linear varying coefficient models (PLVCM) are often considered for analysing longitudinal data a good balance between flexibility and parsimony. The existing estimation variable selection methods this model mainly built upon which subset of variables have or effect on the response is known in advance, say, structure determined. However, application, unreasonable. In work, we propose simultaneous method, can do three types selections: constant effects selection, relevant selection. It...