Najmeh Nakhaei Rad
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Najmeh Nakhaei Rad has a total of 51 co-authors. Here are the top co-authors by number of shared publications:
- Andriette Bekker
Department of Statistics, University of Pretoria, Pretoria, South Africa
7 shared publications - Christophe Ley
University of Luxembourg
4 shared publications - Sphiwe Bonakele Skhosana
Department of Statistics, University of Pretoria, Pretoria, South Africa
4 shared publications - Ding-Geng Chen
College of Health Solutions Arizona State University Phoenix Arizona USA
3 shared publications - Mohammad Arashi
Ferdowsi University of Mashhad
3 shared publications - Vahid Fakoor
Ferdowsi University of Mashhad
3 shared publications
Latest Publications
- Open Access
- 5 November 2026
- Published by Springer Nature in Journal of Statistical Theory and Practice
- Vol. 21 (1) , 10
A Circular Interval-Valued Time Series Model with Application to Wind Direction Prediction
Interval-valued data arise frequently in practice across a wide range of fields, for example, in daily minimum and maximum temperature records and daily opening and closing prices in financial markets. Conventional methods typically aggregate interval observations into single values, such as the mean or midpoint, which may result in a substantial loss of information about the variability and dependency structure. A similar issue arises for circular observations when quantities such as wind direction are recorded as intervals. In Euclidean space, the literature provides only interval-valued autoregressive time series models, whereas on the circle, there remains a clear gap in the modelling of circular interval-valued time series data. The periodic nature of circular data makes standard time series methods inappropriate. To address this gap, this paper introduces an autoregressive time series model for circular interval-valued data. The proposed model accounts for both interval uncertainty and periodic structure. Its performance is evaluated through simulation studies and illustrated using a real dataset, demonstrating its potential as a useful alternative to aggregation-based time series models for the analysis of circular interval data.
- 5 November 2026
- Published by Springer Nature in Journal of Statistical Theory and Practice
- Vol. 21 (1) , 10
Open AccessA Circular Interval-Valued Time Series Model with Application to Wind Direction Prediction
Interval-valued data arise frequently in practice across a wide range of fields, for example, in daily minimum and maximum temperature records and daily opening and closing prices in financial markets. Conventional methods typically aggregate interval observations into single values, such as the mean or midpoint, which may result in a substantial loss of information about the variability and dependency structure. A similar issue arises for circular observations when quantities such as wind direction are recorded as intervals. In Euclidean space, the literature provides only interval-valued autoregressive time series models, whereas on the circle, there remains a clear gap in the modelling of circular interval-valued time series data. The periodic nature of circular data makes standard time series methods inappropriate. To address this gap, this paper introduces an autoregressive time series model for circular interval-valued data. The proposed model accounts for both interval uncertainty and periodic structure. Its performance is evaluated through simulation studies and illustrated using a real dataset, demonstrating its potential as a useful alternative to aggregation-based time series models for the analysis of circular interval data.
- Open Access
- 21 September 2026
- Published by Springer Nature in Theoretical and Applied Climatology
- Vol. 157 (10) , 668
Assessing and correcting the effects of biased sampling on Weibull parameter estimation for wind speed data modelling
In India, routine surface wind observations are generally measured in decimal knots and later disseminated in integer kilometres per hour after operational unit conversion and rounding. Although this reporting procedure is convenient for dissemination, it introduces quantization-induced sampling bias into the released wind speed records. Such discretization can systematically distort the estimation of continuous probability distribution parameters used in probabilistic wind speed modelling. In the present study, the Weibull distribution is adopted as an illustrative framework to quantify this effect and to develop an inverse analytical correction technique for recovering unbiased Weibull shape and scale parameters from disseminated biased estimates. A large Monte Carlo simulation campaign is performed over a wide range of Weibull parameter combinations, and second-order polynomial response surfaces are constructed to map biased estimates back to their unbiased counterparts. Independent out-of-sample validation confirms high reconstruction accuracy with R 2 = 0.9998 for the shape parameter and R 2 = 0.9723 for the scale parameter, accompanied by low RMSE and MAPE values. The practical usefulness of the proposed correction framework is further demonstrated through wind power density estimation over representative Indian wind regimes, where dissemination-induced bias is shown to produce substantial underestimation in low-shape-parameter conditions. The study establishes that operational dissemination-stage rounding, though seemingly minor, can propagate into significant statistical distortion in downstream wind resource assessment unless appropriately corrected.
- 21 September 2026
- Published by Springer Nature in Theoretical and Applied Climatology
- Vol. 157 (10) , 668
Open AccessAssessing and correcting the effects of biased sampling on Weibull parameter estimation for wind speed data modelling
In India, routine surface wind observations are generally measured in decimal knots and later disseminated in integer kilometres per hour after operational unit conversion and rounding. Although this reporting procedure is convenient for dissemination, it introduces quantization-induced sampling bias into the released wind speed records. Such discretization can systematically distort the estimation of continuous probability distribution parameters used in probabilistic wind speed modelling. In the present study, the Weibull distribution is adopted as an illustrative framework to quantify this effect and to develop an inverse analytical correction technique for recovering unbiased Weibull shape and scale parameters from disseminated biased estimates. A large Monte Carlo simulation campaign is performed over a wide range of Weibull parameter combinations, and second-order polynomial response surfaces are constructed to map biased estimates back to their unbiased counterparts. Independent out-of-sample validation confirms high reconstruction accuracy with R 2 = 0.9998 for the shape parameter and R 2 = 0.9723 for the scale parameter, accompanied by low RMSE and MAPE values. The practical usefulness of the proposed correction framework is further demonstrated through wind power density estimation over representative Indian wind regimes, where dissemination-induced bias is shown to produce substantial underestimation in low-shape-parameter conditions. The study establishes that operational dissemination-stage rounding, though seemingly minor, can propagate into significant statistical distortion in downstream wind resource assessment unless appropriately corrected.
- Open Access
- 2 September 2026
- Published by Springer Nature in Journal of Statistical Theory and Practice
- Vol. 20 (4)
Jointly Modeling Dynamic Dependence and Volatility in Natural Resource and Energy Indices Using a Multivariate t-Copula eGARCH Model
This study develops a robust framework for modeling dynamic volatility, asymmetry, and tail dependence in financial returns, focusing on the daily returns of Natural Resource Index () and the Oil and Gas Index (). Standard multivariate volatility models, such as , often fail to adequately capture extreme co-movements, asymmetric dependence, and tail risk under non-normality, potentially producing biased variance and correlation estimates. To address these limitations, we propose a copula-based model with multivariate Student-t () marginals, which jointly captures log-volatility dynamics and flexible tail dependence. Among 18 candidate models, the model with marginals and a multivariate Student-t copula provided the best fit, effectively accounting for heavy tails, volatility clustering, asymmetry, and dynamic dependence. Monte-Carlo simulations based on the empirical estimation confirmed that this framework reduces variance estimation bias, improves coverage probabilities, and more accurately models dynamic correlations and tail risk compared to standard models. This Monte-Carlo study showed that the coverage probability achieved by the model is closer to the target 95% due to ’s ability to capture asymmetric volatility and heavy-tailed behavior in financial returns. Overall, the proposed framework offers superior performance and risk assessment, providing a comprehensive tool for portfolio management, Value-at-Risk estimation, and the analysis of complex dependence structures in financial markets exhibiting asymmetric shocks and heavy-tailed behavior.
- 2 September 2026
- Published by Springer Nature in Journal of Statistical Theory and Practice
- Vol. 20 (4)
Open AccessJointly Modeling Dynamic Dependence and Volatility in Natural Resource and Energy Indices Using a Multivariate t-Copula eGARCH Model
This study develops a robust framework for modeling dynamic volatility, asymmetry, and tail dependence in financial returns, focusing on the daily returns of Natural Resource Index () and the Oil and Gas Index (). Standard multivariate volatility models, such as , often fail to adequately capture extreme co-movements, asymmetric dependence, and tail risk under non-normality, potentially producing biased variance and correlation estimates. To address these limitations, we propose a copula-based model with multivariate Student-t () marginals, which jointly captures log-volatility dynamics and flexible tail dependence. Among 18 candidate models, the model with marginals and a multivariate Student-t copula provided the best fit, effectively accounting for heavy tails, volatility clustering, asymmetry, and dynamic dependence. Monte-Carlo simulations based on the empirical estimation confirmed that this framework reduces variance estimation bias, improves coverage probabilities, and more accurately models dynamic correlations and tail risk compared to standard models. This Monte-Carlo study showed that the coverage probability achieved by the model is closer to the target 95% due to ’s ability to capture asymmetric volatility and heavy-tailed behavior in financial returns. Overall, the proposed framework offers superior performance and risk assessment, providing a comprehensive tool for portfolio management, Value-at-Risk estimation, and the analysis of complex dependence structures in financial markets exhibiting asymmetric shocks and heavy-tailed behavior.
- Open Access
- 22 July 2026
- Published by Springer Nature in Applied Spatial Analysis and Policy
- Vol. 19 (3) , 175
Mapping Hemoglobin Levels Among Women and Children Across Ethiopian Administrative Zones: A Geospatial Small Area Estimation Approach
Hemoglobin, an iron-rich protein in red blood cells, enables oxygen delivery to all body tissues. The hemoglobin records in most countries are collected through surveys. Surveys often provide reliable estimates of target variables for the population at both national and regional administration levels. This study investigates the spatial distribution of hemoglobin levels among women and children across Ethiopian administrative zones. Because survey sample sizes are often insufficient to produce reliable direct estimates at the zonal level, small area estimation methods are used to improve the precision of estimates. Small area estimation provides reliable estimates at the local zones by combining survey, census, and spatial datasets. Therefore, we focus on geospatial small area estimation to provide reliable and precise hemoglobin estimates for women and children. Therefore, our prime objective is to provide reliable and precise estimates of hemoglobin levels for women and children using geospatial small area estimation by integrating the survey and census datasets. An area-level Fay-Herriot modeling framework was implemented to estimate hemoglobin levels across the studied zones. The results show that for both children and women, the Spatial Empirical Best Linear Unbiased Predictor (SEBLUP) consistently produces the lowest coefficients of variation (CVs), followed by the standard Empirical Best Linear Unbiased Predictor (EBLUP), while direct survey estimates display the highest variability. The comparison of CVs for hemoglobin estimates demonstrates that children’s mean CVs decrease from 13.64 (Direct survey estimates) to 11.29 (EBLUP) and then to 6.06 (SEBLUP). Similarly, women’s mean CVs decline from 13.48 (Direct survey estimates) to 10.71 (EBLUP) and then 6.37 (SEBLUP), confirming that SEBLUP provides the highest precision for both children and women, consistently producing the lowest CVs across all statistical measures. The results in the maps indicate the spatial disparities in hemoglobin levels for both women and children among the Ethiopian administrative zones. Policymakers will significantly benefit from these local area-level estimates, and researchers will gain methodological insights into the application of geospatial small-area estimation.
- 22 July 2026
- Published by Springer Nature in Applied Spatial Analysis and Policy
- Vol. 19 (3) , 175
Open AccessMapping Hemoglobin Levels Among Women and Children Across Ethiopian Administrative Zones: A Geospatial Small Area Estimation Approach
Hemoglobin, an iron-rich protein in red blood cells, enables oxygen delivery to all body tissues. The hemoglobin records in most countries are collected through surveys. Surveys often provide reliable estimates of target variables for the population at both national and regional administration levels. This study investigates the spatial distribution of hemoglobin levels among women and children across Ethiopian administrative zones. Because survey sample sizes are often insufficient to produce reliable direct estimates at the zonal level, small area estimation methods are used to improve the precision of estimates. Small area estimation provides reliable estimates at the local zones by combining survey, census, and spatial datasets. Therefore, we focus on geospatial small area estimation to provide reliable and precise hemoglobin estimates for women and children. Therefore, our prime objective is to provide reliable and precise estimates of hemoglobin levels for women and children using geospatial small area estimation by integrating the survey and census datasets. An area-level Fay-Herriot modeling framework was implemented to estimate hemoglobin levels across the studied zones. The results show that for both children and women, the Spatial Empirical Best Linear Unbiased Predictor (SEBLUP) consistently produces the lowest coefficients of variation (CVs), followed by the standard Empirical Best Linear Unbiased Predictor (EBLUP), while direct survey estimates display the highest variability. The comparison of CVs for hemoglobin estimates demonstrates that children’s mean CVs decrease from 13.64 (Direct survey estimates) to 11.29 (EBLUP) and then to 6.06 (SEBLUP). Similarly, women’s mean CVs decline from 13.48 (Direct survey estimates) to 10.71 (EBLUP) and then 6.37 (SEBLUP), confirming that SEBLUP provides the highest precision for both children and women, consistently producing the lowest CVs across all statistical measures. The results in the maps indicate the spatial disparities in hemoglobin levels for both women and children among the Ethiopian administrative zones. Policymakers will significantly benefit from these local area-level estimates, and researchers will gain methodological insights into the application of geospatial small-area estimation.
- Open Access
- 10 July 2026
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 35 (8) , 1746-1784
Deep learning embedded latent class joint modelling of time-to-event and longitudinal data
Joint modelling of longitudinal and time-to-event data is a powerful tool for analysing complex medical data. Joint latent class models (JLCMs), in particular, provide additional prognostic insight by clustering subjects into latent subgroups within a mixture of experts framework. Traditionally, the mixing proportions are modelled using a multinomial logistic regression. Recent work in mixture regression has proposed a neural network alternative, implemented through the neural network expectation–maximization algorithm, which offers greater flexibility by capturing nonlinearities that may be missed by logistic regression. In this article, we extend the JLCM by applying the neural network to the modelling the mixing proportions captured in the class-membership submodel. The model estimated using an expectation–maximization-type algorithm. We call this the neural-network JLCM (NN-JLCM). We formally define the method, evaluate its performance through extensive simulations, and apply it to real data. The results show that NN-JLCM improves predictive accuracy and class separation compared with traditional JLCMs, while maintaining competitiveness in terms of overall model fit. The code for the article is available at https://github.com/tharris0924/nn-jlcm_p1 .
- 10 July 2026
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 35 (8) , 1746-1784
Open AccessDeep learning embedded latent class joint modelling of time-to-event and longitudinal data
Joint modelling of longitudinal and time-to-event data is a powerful tool for analysing complex medical data. Joint latent class models (JLCMs), in particular, provide additional prognostic insight by clustering subjects into latent subgroups within a mixture of experts framework. Traditionally, the mixing proportions are modelled using a multinomial logistic regression. Recent work in mixture regression has proposed a neural network alternative, implemented through the neural network expectation–maximization algorithm, which offers greater flexibility by capturing nonlinearities that may be missed by logistic regression. In this article, we extend the JLCM by applying the neural network to the modelling the mixing proportions captured in the class-membership submodel. The model estimated using an expectation–maximization-type algorithm. We call this the neural-network JLCM (NN-JLCM). We formally define the method, evaluate its performance through extensive simulations, and apply it to real data. The results show that NN-JLCM improves predictive accuracy and class separation compared with traditional JLCMs, while maintaining competitiveness in terms of overall model fit. The code for the article is available at https://github.com/tharris0924/nn-jlcm_p1 .
- Open Access
- 10 May 2026
- Published by Wiley in Environmetrics
- Vol. 37 (4)
Predicting Direction of High‐Speed Wind: A Bayesian Approach to Conditional Cylindrical Distributions
High‐speed winds pose several dangers to both the built environment and the natural environment, affecting human lives and livelihoods, as well as endangering wildlife. On the other hand, high‐speed winds can be beneficially harnessed through the optimal orientation of turbines to maximize wind energy production. The management response to high‐speed winds is a complex process because not only is their speed of interest, but even more so, their direction. High‐speed winds are generally seasonal. In this paper, we propose a Bayesian prediction model for a circular variable (such as wind direction) conditional on a linear variable (such as high‐speed wind), model that can account for any sparsity in the (circular, linear) pair of data. The proposed Bayesian method comprises a mixture of conditional cylindrical models to capture potential multimodality in the distribution of the circular component and compares two different cylindrical distributions, namely the Abe–Ley distribution and the Kalaylioglu distribution. Monte Carlo simulation studies conducted for both distributions establish the benefit of our proposed method. Finally, we demonstrate the developed posterior predictive distribution in a real application, where we predict the direction of high‐speed winds using hourly in situ wind data from a location in South Africa.
- 10 May 2026
- Published by Wiley in Environmetrics
- Vol. 37 (4)
Open AccessPredicting Direction of High‐Speed Wind: A Bayesian Approach to Conditional Cylindrical Distributions
High‐speed winds pose several dangers to both the built environment and the natural environment, affecting human lives and livelihoods, as well as endangering wildlife. On the other hand, high‐speed winds can be beneficially harnessed through the optimal orientation of turbines to maximize wind energy production. The management response to high‐speed winds is a complex process because not only is their speed of interest, but even more so, their direction. High‐speed winds are generally seasonal. In this paper, we propose a Bayesian prediction model for a circular variable (such as wind direction) conditional on a linear variable (such as high‐speed wind), model that can account for any sparsity in the (circular, linear) pair of data. The proposed Bayesian method comprises a mixture of conditional cylindrical models to capture potential multimodality in the distribution of the circular component and compares two different cylindrical distributions, namely the Abe–Ley distribution and the Kalaylioglu distribution. Monte Carlo simulation studies conducted for both distributions establish the benefit of our proposed method. Finally, we demonstrate the developed posterior predictive distribution in a real application, where we predict the direction of high‐speed winds using hourly in situ wind data from a location in South Africa.
- Open Access
- 28 March 2026
- Published by Springer Nature in Advances in Data Analysis and Classification
- p. 1-30
Model-based clustering using a new mixture of circular regressions
Regression models, where the response variable is circular, are common in areas such as biology, geology and meteorology. A typical model assumes that the conditional distribution of the response follows a von-Mises distribution. However, this assumption is inadequate when the response variable is multimodal. For this reason, in this paper, a finite mixture of regressions model is proposed for the case of a circular response variable and a set of circular and/or linear covariates. Mixture models are very useful when the underlying population is multimodal. Despite the prevalence of multimodality in regression modelling of circular data, the use of mixtures of regressions has received no attention in the literature. This paper aims to close this knowledge gap. To estimate the proposed model, we develop a maximum likelihood estimation procedure via the Expectation-Maximization algorithm. An extensive simulation study is used to demonstrate the practical use and performance of the proposed model and estimation procedure. In addition, the model is shown to be useful as a model-based clustering tool. Lastly, the model is applied to a real dataset from a wind farm in South Africa which highlights its advantage compared to a competitive approach for model-based clustering of circular data.
- 28 March 2026
- Published by Springer Nature in Advances in Data Analysis and Classification
- p. 1-30
Open AccessModel-based clustering using a new mixture of circular regressions
Regression models, where the response variable is circular, are common in areas such as biology, geology and meteorology. A typical model assumes that the conditional distribution of the response follows a von-Mises distribution. However, this assumption is inadequate when the response variable is multimodal. For this reason, in this paper, a finite mixture of regressions model is proposed for the case of a circular response variable and a set of circular and/or linear covariates. Mixture models are very useful when the underlying population is multimodal. Despite the prevalence of multimodality in regression modelling of circular data, the use of mixtures of regressions has received no attention in the literature. This paper aims to close this knowledge gap. To estimate the proposed model, we develop a maximum likelihood estimation procedure via the Expectation-Maximization algorithm. An extensive simulation study is used to demonstrate the practical use and performance of the proposed model and estimation procedure. In addition, the model is shown to be useful as a model-based clustering tool. Lastly, the model is applied to a real dataset from a wind farm in South Africa which highlights its advantage compared to a competitive approach for model-based clustering of circular data.
- Open Access
Robust semi-parametric mixtures of linear experts using the contaminated Gaussian distribution
Semi- and non-parametric mixture of regressions are a very useful flexible class of mixture of regressions in which some or all of the parameters are non-parametric functions of the covariates. These models are, however, based on the Gaussian assumption of the component error distributions. Thus, their estimation is sensitive to outliers and heavy-tailed error distributions. In this paper, we propose semi- and non-parametric contaminated Gaussian mixture of regressions to robustly estimate the parametric and/or non-parametric terms of the models in the presence of mild outliers. The virtue of using a contaminated Gaussian error distribution is that we can simultaneously perform model-based clustering of observations and model-based outlier detection. We propose two algorithms, an expectation-maximization (EM)-type algorithm and an expectation-conditional-maximization (ECM)-type algorithm, to perform maximum likelihood and local-likelihood kernel estimation of the parametric and non-parametric of the proposed models, respectively. The robustness of the proposed models is examined using an extensive simulation study. The practical utility of the proposed models is demonstrated using real data.
Open AccessRobust semi-parametric mixtures of linear experts using the contaminated Gaussian distribution
Semi- and non-parametric mixture of regressions are a very useful flexible class of mixture of regressions in which some or all of the parameters are non-parametric functions of the covariates. These models are, however, based on the Gaussian assumption of the component error distributions. Thus, their estimation is sensitive to outliers and heavy-tailed error distributions. In this paper, we propose semi- and non-parametric contaminated Gaussian mixture of regressions to robustly estimate the parametric and/or non-parametric terms of the models in the presence of mild outliers. The virtue of using a contaminated Gaussian error distribution is that we can simultaneously perform model-based clustering of observations and model-based outlier detection. We propose two algorithms, an expectation-maximization (EM)-type algorithm and an expectation-conditional-maximization (ECM)-type algorithm, to perform maximum likelihood and local-likelihood kernel estimation of the parametric and non-parametric of the proposed models, respectively. The robustness of the proposed models is examined using an extensive simulation study. The practical utility of the proposed models is demonstrated using real data.
- Open Access
Model-based clustering using a new mixture of circular regressions
Regression models, where the response variable is circular, are common in areas such as biology, geology and meteorology. A typical model assumes that the conditional distribution of the response follows a von-Mises distribution. However, this assumption is inadequate when the response variable is multimodal. For this reason, in this paper, a finite mixture of regressions model is proposed for the case of a circular response variable and a set of circular and/or linear covariates. Mixture models are very useful when the underlying population is multimodal. Despite the prevalence of multimodality in regression modelling of circular data, the use of mixtures of regressions has received no attention in the literature. This paper aims to close this knowledge gap. To estimate the proposed model, we develop a maximum likelihood estimation procedure via the Expectation-Maximization algorithm. An extensive simulation study is used to demonstrate the practical use and performance of the proposed model and estimation procedure. In addition, the model is shown to be useful as a model-based clustering tool. Lastly, the model is applied to a real dataset from a wind farm in South Africa.
Open AccessModel-based clustering using a new mixture of circular regressions
Regression models, where the response variable is circular, are common in areas such as biology, geology and meteorology. A typical model assumes that the conditional distribution of the response follows a von-Mises distribution. However, this assumption is inadequate when the response variable is multimodal. For this reason, in this paper, a finite mixture of regressions model is proposed for the case of a circular response variable and a set of circular and/or linear covariates. Mixture models are very useful when the underlying population is multimodal. Despite the prevalence of multimodality in regression modelling of circular data, the use of mixtures of regressions has received no attention in the literature. This paper aims to close this knowledge gap. To estimate the proposed model, we develop a maximum likelihood estimation procedure via the Expectation-Maximization algorithm. An extensive simulation study is used to demonstrate the practical use and performance of the proposed model and estimation procedure. In addition, the model is shown to be useful as a model-based clustering tool. Lastly, the model is applied to a real dataset from a wind farm in South Africa.
- 8 December 2025
- Published by Springer Nature in Journal of Statistical Theory and Practice
- Vol. 20 (1) , 10
Copula-Based Vector Autoregressive Modeling for Volatility Index Data
- 8 December 2025
- Published by Springer Nature in Journal of Statistical Theory and Practice
- Vol. 20 (1) , 10
Copula-Based Vector Autoregressive Modeling for Volatility Index Data