Yazd UniversityJournal of Statistical Modelling: Theory and Applications2676-73927120260908Bayesian multiple change-point detection: A methodological comparison of exact dynamic programming and nonparametric hidden Markov approaches119424310.22034/jsmta.2026.23821.1200ENMasoudMajidizadehDepartment of Statistics, Faculty of Mathematical Sciences, Shahid Beheshti University, Tehran, Iran0009-0002-1626-7693Journal Article20251016Detecting structural breaks is crucial for understanding time series dynamics, particularly in highly volatile financial markets. This study systematically compares two prominent Bayesian change-point methodologies: an exact dynamic programming approach that eliminates Markov chain Monte Carlo sampling, and a Dirichlet process hidden Markov model that flexibly infers latent regime counts and persistence. Analyzing daily S&P 500 returns spanning 2000 to 2024-a period encompassing major upheavals like the global financial crisis and the COVID-19 pandemic-we evaluate both models across detection accuracy, parameter estimation, computational efficiency, and predictive performance. Despite fundamentally different assumptions, the models exhibit striking concordance, identifying major macroeconomic shifts with timing differences of mere days and parameter variations of under 0.01%. However, a clear trade-off emerges: the exact method is approximately 90 times faster, whereas the Dirichlet process hidden Markov model provides marginally superior predictive accuracy. These insights offer actionable guidelines for selecting the optimal framework based on computational and predictive priorities.https://jsm.yazd.ac.ir/article_4243_962f89b51a155fdb7f4cfb6252bd5143.pdfYazd UniversityJournal of Statistical Modelling: Theory and Applications2676-73927120260912Generalization of reliability in the odd Lindley-G family2135425110.22034/jsmta.2026.24914.1229ENMaryamDarijaniDepartment of Statistics, Higher Education Complex of Bam, Kerman, IranHojatollahZakerzadehDepartment of Statistics, Yazd University, Yazd, IranJournal Article20260712This study presents a generalized numerical framework for evaluating the reliability criterion in the odd Lindley-G family of distributions. Although the odd Lindley-G family provides a flexible mechanism for extending classical baseline models, the existing analytical formulation for reliability estimation suffers from two critical limitations: it is restricted to cases where the strength and stress variables share an identical baseline distribution, and its series expansion diverges even under homogeneous baseline conditions. To address these limitations, we develop a robust numerical integration framework that accurately computes reliability for arbitrary and distinct baseline distributions within the odd Lindley-G family. The proposed method is validated through an extensive simulation study and a practical engineering application involving strength and stress data with different baseline distributions. Our findings confirm that the numerical framework not only generalizes the applicability of the odd Lindley-G family but also delivers a stable, accurate, and computationally efficient tool for advanced reliability analysis.https://jsm.yazd.ac.ir/article_4251_abbff5dcb23fc491186988f649a1d3b6.pdfYazd UniversityJournal of Statistical Modelling: Theory and Applications2676-73927120260101Simulation-based path analysis for qualitative mediators: A continuous proxy approach3753427310.22034/jsmta.2026.24305.1211ENAlirezaPakgoharDepartent of Statistics, Payame Noor University, Tehran, Iran0000-0003-2512-1581MohadesehKhaliliDepartment of Engineering, University of Ardakan, Ardakan, IranJournal Article20260208Conventional path analysis relies on ordinary least squares assumptions, requiring continuous variables. When mediators are qualitative, researchers typically resort to logistic regression, creating a scale incompatibility problem in which coefficients cannot be directly multiplied to estimate indirect effects. This study proposes a Categorical-to-Continuous Signal Transformation technique to address this limitation. Unlike standard dummy coding, we introduce a simulation-based algorithm that generates a continuous proxy variable for the qualitative mediator. By extracting the probabilistic centroid of each category using the group mean and standard error of the dependent variable, we simulate a distribution that retains the structural signal while satisfying ordinary least squares normality assumptions. The method was validated using both primary and synthetic datasets. Results demonstrate that the simulated proxy variables successfully recover the directional pathways and maintain statistical significance consistent with classical analysis of variance. The approach effectively captures the latent influence of categories without the bias of individual-level noise. The proposed simulation method offers a robust alternative for mediation analysis involving qualitative variables. By bridging the gap between categorical data and linear modeling, it enables the direct calculation of indirect effects within a unified ordinary least squares framework, eliminating the need for complex non-linear approximations.https://jsm.yazd.ac.ir/article_4273_66ce5822f0c1d881db978154311596ca.pdfYazd UniversityJournal of Statistical Modelling: Theory and Applications2676-73927120260101Estimation and prediction for the XLindley distribution under progressively Type-II censoring5585428010.22034/jsmta.2026.24546.1217ENRoshanakZamanDepartment of Statistics, Payame Noor University, Tehran, Iran0000-0001-9885-5311AdelehFallahDepartment of Statistics, Payame Noor University, Tehran, Iran0000-0002-9849-6037Journal Article20260509In this paper, we consider the problems of estimating unknown parameters as well as predicting the failure times of the removed units in multiple stages of the progressively censored sample coming from an XLindley distribution. We obtain maximum likelihood, moment-based, bootstrap, and Bayes estimates under the squared error loss function. Since the integrals associated with the Bayesian estimators do not admit closed-form solutions, Lindley's approximation, the Markov chain Monte Carlo method, and Tierney and Kadane's approximation are used to approximate these integrals. We also compute the confidence intervals based on the asymptotic distribution of the maximum likelihood estimate, the confidence intervals based on the asymptotic distribution of the logarithm of the maximum likelihood estimate, the exact confidence intervals, the confidence intervals based on the likelihood ratio statistic, the bootstrap confidence intervals, and the Bayesian confidence intervals. Different point and interval predictors are derived based on classical and Bayesian approaches. A real example is provided to further explain the methods introduced. A Monte Carlo simulation study is conducted to evaluate and compare the performance of different estimation and prediction methods.https://jsm.yazd.ac.ir/article_4280_6815ff46316219f67359f52c65556e2e.pdfYazd UniversityJournal of Statistical Modelling: Theory and Applications2676-73927120260101A tale of two curves: Exploring the unexpected relationship between the Lorenz curve and the ROC curve87111429110.22034/jsmta.2026.24527.1216ENZahraBehdaniDepartment of Mathematics and Statistics, Faculty of Energy and Data Sciences, Behbahan Khatam Alanbia University of Technology, Behbahan, IranVahidehAhrariDepartment of Computer Science, Faculty of Mathematical Sciences, Shahrekord University, Shahrekord, Iran0000-0002-5628-3893Gholam RezaMohtashami BorzadaranDepartment of Statistics, Faculty of Mathematical Sciences, Ferdowsi University of Mashhad, IranJournal Article20260505This paper provides a conceptual and methodological synthesis of the relationship between two tools that originate from different disciplines: the Lorenz curve, widely used to study economic inequality, and the receiver operating characteristic curve, widely used to evaluate binary classification and predictive discrimination. Building on established results linking the Gini coefficient to the area under the receiver operating characteristic curve, the study examines their relationship from a broader ranking-based and distribution-based perspective, clarifying their mathematical connections, distributional foundations, and distinct interpretations through analytical examples, Monte Carlo simulations, and applied evaluations under different sample sizes and class imbalance conditions. By clarifying these connections and examining their empirical behavior beyond idealized settings, the study provides a cross-disciplinary framework for interpreting ranking-based measures in both inequality measurement and predictive classification.https://jsm.yazd.ac.ir/article_4291_3253f205d06e1d8756bf3324c7ee1cc1.pdfYazd UniversityJournal of Statistical Modelling: Theory and Applications2676-73927120261005Clipped self-normalized Max-CUSUM for sparse mean breaks in high-dimensional heavy-tailed locally stationary time series113127429210.22034/jsmta.2026.24764.1222ENLeilaMarvian MashhadImam Reza International University, Mashhad, Iran0009-0009-1620-5346Journal Article20260616We develop a robust procedure for detecting and locating a single sparse mean break in high-dimensional, locally stationary time series with heavy-tailed innovations and intermittent contamination. The method combines coordinatewise clipping, local self-normalization, and a maximum cumulative sum scan. In contrast with an oracle calibration, the proposed implementation uses a locally centered block-multiplier bootstrap, so it does not require a known covariance structure. We establish negligible clipping bias under finite moments, a Gaussian approximation for the scan statistic, bootstrap validity under the null, a sparse detection condition, and a localization rate. The theory is accompanied by sufficient conditions for time-varying linear and nonlinear filters. A fully specified Monte Carlo study uses 250 replications per main design, reports Wilson confidence intervals and Monte Carlo standard errors, compares oracle and data-driven calibration, and examines sensitivity to clipping, window, ridge, scanning, dimension, and sparsity. The proposed method is competitive with raw max scanning in clean settings and is particularly advantageous for sparse breaks under heavy tails and contamination; dense ℓ<sub>2</sub> aggregation need not be dominated in clean, dense-signal regimes.https://jsm.yazd.ac.ir/article_4292_286fb673d540ec108e8fdb7984389edf.pdf