Journal of Statistical Sciences

Stochastic Comparison of Extreme Order Statistics Corresponding to Two Sets of Dependent Scale Additive Hazard Rate Model Random Variables under Random Shocks

Marzieh Shekari

In this article, we investigate stochastic comparisons of the lifetimes of series and parallel systems comprising components following the scale-additive hazard rate model, subject to random shocks, where dependence among component lifetimes is modeled via Archimedean copulas. By imposing suitable conditions on the baseline distribution, model parameters, generator functions, and shock occurrence probabilities, we establish sufficient criteria for comparing the lifetimes of two systems under both the usual stochastic order and the hazard rate order. The results demonstrate how parameter heterogeneity and dependence structure simultaneously influence system reliability. Several numerical examples are also provided to substantiate the theoretical findings.

Estimation and Prediction in Lindley-Weibel Distribution Based on Record Data

Ali Akbar Heydari

The aim of this paper is to develop statistical inference methods for the Lindley-Weibull distribution when only upper record values are available. Using record theory, likelihood functions for parameter estimation are derived, and maximum likelihood estimators are presented. Additionally, a method for predicting future records based on observed records is proposed. The performance of the methods is evaluated through a simulation study and an application to real flood discharge data. The results indicate that the Lindley-Weibull distribution has high flexibility in modeling record data, and the proposed inference methods have appropriate accuracy.

Comparative Analysis of Parameter Estimation Methods for Periodic Autoregressive Models with Scale Mixture of Skew Normal Distributions in the Face of Outliers and Heavy Tailed Data

Tahere Manouchehri

In this paper, we present a comprehensive review and comparative analysis of estimation methods for periodic autoregressive (PAR) models driven by scale mixture of skew-normal (SMSN) innovations, a flexible class suitable for modeling both symmetric and asymmetric data. Expectation-conditional maximization algorithms are employed to develop maximum likelihood, maximum a posteriori, and Bayesian estimation procedures. A thorough evaluation of these methods is conducted using simulation studies, with particular attention to asymptotic properties and robustness against outliers, high peaks, and heavy tails. To demonstrate their practical utility, these methods are applied to monthly Google stock price data.

Bayesian Analysis of Skewed Spatio-Temporal Data Using Bayesian Neural Fileld and Variational Inference

Fatemeh Hosseini

Spatio-temporal data often exhibit complex dependence structures and skewness, which makes their modeling with classical frameworks, such as Gaussian random fields, either computationally expensive or overly restrictive. In this paper, we introduce a novel Bayesian Neural Field framework for modeling skewed spatio-temporal processes. The proposed approach incorporates spatial and temporal coordinates, along with explanatory variables and prior distributions, allowing flexible representation of dependence and skewness, as well as prediction at new locations and at unseen time points. Parameter inference is performed using variational inference, which offers both computational efficiency and the ability to quantify uncertainty. Simulation results demonstrate that the proposed framework achieves higher accuracy and faster computation compared to standard Monte Carlo methods.

A Bayesian Approach and Vecchia Grouping for Nonparametric Estimation of Nonstationary Spatial Covariance in Big Data

Ali Mohammadian Mosammam

This paper presents a nonparametric Bayesian method for estimating nonstationary covariance structures in big spatial datasets. The approach extends the Vecchia approximation and assumes conditional independence among ordered data points, leading to a sparse precision matrix and sparse Cholesky decomposition. This enables modeling an $n$-dimensional Gaussian process as a sequence of Bayesian linear regressions. Data ordering via maximum minimum distance improves model performance. Applying the grouping algorithm to ordered data removes weak dependencies and defines a block-sparse covariance structure, significantly reducing computational burden and enhancing accuracy. Simulations and real data analysis show that posterior samples from the proposed method yield narrower uncertainty intervals than those from ungrouped approaches.

A Study of the Complexity of Conditional Mixed Systems and Their Duals with Gamma Components Using the Jensen Cumulative Residual Extropy Divergence Index

Reza Alizadeh Noughabi

In this paper, a novel index entitled the Jensen cumulative residual extropy divergence is investigated for the analysis and measurement of the behavioral complexity of conditional mixed systems. First, using the vector of conditional coefficients obtained from the signature vector, the behavior of this measure is analytically examined for a class of coherent systems as well as their dual systems, in the case where the components follow gamma distributions. Then, simulations are performed to evaluate the obtained results. The results of this paper show that the minimum complexity is achieved by coherent $k$-out-of-$n$ systems with the Jensen cumulative residual extropy divergence equal to zero. Moreover, the results indicate that duality of systems does not necessarily lead to equality of the Jensen cumulative residual extropy divergence in conditional mixed systems; rather, this index is sensitive to component weighting, order statistics, and the structural interaction among the components of the system.

INAR(1)-PK: ‎Modeling Discrete-Time Count Data with Overdispersion

mahdi rasekhi

In this paper, a first-order integer-valued autoregressive process with non-negative integer values is introduced, based on the binomial thinning operator and driven by Poisson-Komal distributed noise. To estimate the parameters of the proposed model, two estimation methods are investigated: Conditional Maximum Likelihood Estimation and the Yule–Walker Method. Furthermore, the performance of these estimation techniques is evaluated through a simulation study. In addition, the practical applicability of the proposed model is demonstrated using two real-world datasets from the field of veterinary sciences.

Entropy and Extropy of the Wavelet Energy Distribution in Reciprocal Fractal Systems

Alireza Pakgohar

This study investigates the wavelet energy distribution in high-frequency fractal systems and analyzes its characteristics using information-theoretic measures. The main innovation of this paper lies in modeling the wavelet energy distribution ($p_j$) using a truncated geometric distribution and incorporating the concept of extropy to quantify system complexity. It is demonstrated that this distribution is strongly influenced by the fractal parameter $alpha$ and the number of decomposition levels $M$. By computing wavelet entropy and extropy as measures of disorder and information, respectively—the study provides a quantitative analysis of the complexity of these systems. The paper further examines key properties of this distribution, including its convergence to geometric, uniform, and degenerate distributions under limiting conditions (e.g., $M to infty$ or $alpha to 0$). Results indicate that entropy and extropy serve as complementary tools for a comprehensive description of system behavior: while entropy measures disorder, extropy reflects the degree of information and certainty. This approach establishes a novel framework for analyzing real-world signals with varying parameters and holds potential applications in the analysis of fractal signals and modeling of complex systems in fields such as finance and biology.

To validate the theoretical findings, synthetic fractal signals (fractional Brownian motion) with varying fractal parameters ($alpha$) and decomposition levels ($M$) were simulated. Numerical results show that wavelet entropy increases significantly with the number of decomposition levels ($M$), whereas extropy exhibits slower growth and saturates at higher decomposition levels. These findings underscore the importance of selecting an appropriate decomposition level. The proposed combined framework offers a powerful tool for analyzing and modeling complex, non-stationary systems in domains such as finance and biology.