Hossein Safari
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
Research Subjects
Research Interests
Hossein Safari has a total of 310 co-authors. Here are the top co-authors by number of shared publications:
- Nasibe Alipour
University of Guilan Department of Physics 41335-1914 Rasht
16 shared publications - Nastaran Farhang
Sydney Institute for Astronomy, University of Sydney , NSW 2006
8 shared publications - Michael. S. Wheatland
University of Sydney
6 shared publications - Hamed Ghasemi
Department of Physics, Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan 45137-66731, Iran
5 shared publications S. Nasiri
University of Zanjan
5 shared publications- Abolfazl Dinmohammadi
University of Zanjan
4 shared publications
Latest Publications
- 26 October 2026
- Published by Springer Nature in Astrophysics
- p. 1-11
Magnetic polarity distribution and evolution of an active region moss areas; a case study
- 26 October 2026
- Published by Springer Nature in Astrophysics
- p. 1-11
Magnetic polarity distribution and evolution of an active region moss areas; a case study
- 1 September 2026
- Published by Springer Nature in Astrophysics
- p. 1-18
Dependency of flare radiations to magnetic flux variations
- 1 September 2026
- Published by Springer Nature in Astrophysics
- p. 1-18
Dependency of flare radiations to magnetic flux variations
- Open Access
- 24 April 2026
- Published by American Astronomical Society in The Astronomical Journal
- Vol. 171 (5) , 313
Ensemble Machine Learning Approach to Estimate the Asteroseismic Indices for δ Scuti Stars Observed by TESS
We develop an ensemble machine learning framework to estimate key asteroseismic parameters—namely the frequency of the highest peak ( ν ( A max )), the frequency of maximum oscillation power ( ν max ), and the large frequency separation (Δ ν )—directly from TESS light curves of δ Scuti stars ( δ Sct stars). For each light curve, we extract a set of features comprising statistical moments, Principal Component Analysis, autocorrelation function, spectral features from the fast Fourier transform, and multiscale features from the discrete wavelet transform. These features are used to train a stacked regressor model composed of Random Forest, Gradient Boosting, and Ridge Regressors. We train and evaluate the model using 583 δ Sct stars randomly selected from a total of 643 samples, repeatedly applying random 80/20 splits across 100 iterations. The model achieves high predictive accuracy with R 2 scores exceeding 0.77 for all asteroseismic quantities. We validate generalizability by applying the trained model to the remaining 60 δ Sct stars, not seen during training. The predicted values show strong agreement with traditional asteroseismic measurements, confirming the effectiveness of this framework for large-scale, automated asteroseismic analysis. Furthermore, the proposed framework allows for the estimation of asteroseismic indices across new 251 δ Sct stars.
- 24 April 2026
- Published by American Astronomical Society in The Astronomical Journal
- Vol. 171 (5) , 313
Open AccessEnsemble Machine Learning Approach to Estimate the Asteroseismic Indices for δ Scuti Stars Observed by TESS
We develop an ensemble machine learning framework to estimate key asteroseismic parameters—namely the frequency of the highest peak ( ν ( A max )), the frequency of maximum oscillation power ( ν max ), and the large frequency separation (Δ ν )—directly from TESS light curves of δ Scuti stars ( δ Sct stars). For each light curve, we extract a set of features comprising statistical moments, Principal Component Analysis, autocorrelation function, spectral features from the fast Fourier transform, and multiscale features from the discrete wavelet transform. These features are used to train a stacked regressor model composed of Random Forest, Gradient Boosting, and Ridge Regressors. We train and evaluate the model using 583 δ Sct stars randomly selected from a total of 643 samples, repeatedly applying random 80/20 splits across 100 iterations. The model achieves high predictive accuracy with R 2 scores exceeding 0.77 for all asteroseismic quantities. We validate generalizability by applying the trained model to the remaining 60 δ Sct stars, not seen during training. The predicted values show strong agreement with traditional asteroseismic measurements, confirming the effectiveness of this framework for large-scale, automated asteroseismic analysis. Furthermore, the proposed framework allows for the estimation of asteroseismic indices across new 251 δ Sct stars.
- Open Access
- 23 January 2026
- Published by American Astronomical Society in The Astrophysical Journal
- Vol. 997 (2) , 209
Modeling of IRIS Spectral Line Profiles via Markov Chain Monte Carlo in Bayesian Approach
Spectral line analysis is essential for investigating the solar atmospheric dynamic processes. Modeling the line profiles provides key insights into plasma motions and variations in temperature and density. This study aims to enhance the parameter estimation of spectral line profiles observed by the Interface Region Imaging Spectrograph through the application of an ensemble approach within the Bayesian Markov Chain Monte Carlo (MCMC) framework. We analyze silicon ion lines (Si IV 1394 and 1403 Å) for samples in the plage, flaring/nonflaring active regions, coronal hole, and boundary of coronal hole and quiet Sun, fitting them with single-Gaussian, double-Gaussian, and Voigt models using the Bayesian MCMC approach. The performance of the model is assessed using statistical metrics, including the reduced chi-square ( χ 2 ), Akaike information criterion, and Bayesian information criterion. The Bayesian MCMC method not only provides precise parameter estimates with associated uncertainties but also converges efficiently when initialized with Levenberg–Marquardt least-squares results. Overall, this approach enhances the accuracy of spectral line modeling and model selection, contributing to a more detailed understanding of solar atmospheric dynamics.
- 23 January 2026
- Published by American Astronomical Society in The Astrophysical Journal
- Vol. 997 (2) , 209
Open AccessModeling of IRIS Spectral Line Profiles via Markov Chain Monte Carlo in Bayesian Approach
Spectral line analysis is essential for investigating the solar atmospheric dynamic processes. Modeling the line profiles provides key insights into plasma motions and variations in temperature and density. This study aims to enhance the parameter estimation of spectral line profiles observed by the Interface Region Imaging Spectrograph through the application of an ensemble approach within the Bayesian Markov Chain Monte Carlo (MCMC) framework. We analyze silicon ion lines (Si IV 1394 and 1403 Å) for samples in the plage, flaring/nonflaring active regions, coronal hole, and boundary of coronal hole and quiet Sun, fitting them with single-Gaussian, double-Gaussian, and Voigt models using the Bayesian MCMC approach. The performance of the model is assessed using statistical metrics, including the reduced chi-square ( χ 2 ), Akaike information criterion, and Bayesian information criterion. The Bayesian MCMC method not only provides precise parameter estimates with associated uncertainties but also converges efficiently when initialized with Levenberg–Marquardt least-squares results. Overall, this approach enhances the accuracy of spectral line modeling and model selection, contributing to a more detailed understanding of solar atmospheric dynamics.
- Open Access
- 12 September 2025
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society: Letters
- Vol. 544 (1) , L1-L7
Machine learning based scaling relations for mass, radius, and age of Solar-like oscillating stars
We investigate scaling relations for stellar mass, radius, and age in stars exhibiting solar-like oscillations using two observational data sets comprising 110 stars. Using data set 1 (80 stars), we derive empirical scaling relations incorporating asteroseismic parameters – the frequency of maximum oscillation power, large, and small frequency separations – alongside effective temperature and metallicity. To account for non-Gaussian uncertainties in the stellar parameters, we employ an ensemble machine learning approach that combines Ridge, Huber, and Gradient Boosting regressors to predict stellar mass, radius, and age, along with deriving the corresponding coefficients for the scaling relations. We assess coefficient uncertainties by combining jackknife and bootstrap resampling techniques. Using the resulting scaling relations, we predict the mass, radius, and age of 30 stars in data set 2. These predictions closely match observational data, even when the parameters slightly differ from those in the training set. The predicted stellar ages range from 2.29 to 10.09 Gyr, consistent with Solar-like oscillators. Our findings show that combining machine learning regression with resampling-based uncertainty estimation provides an effective and reliable approach for deriving scaling relations, offering valuable insights for stellar characterization in asteroseismic research.
- 12 September 2025
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society: Letters
- Vol. 544 (1) , L1-L7
Open AccessMachine learning based scaling relations for mass, radius, and age of Solar-like oscillating stars
We investigate scaling relations for stellar mass, radius, and age in stars exhibiting solar-like oscillations using two observational data sets comprising 110 stars. Using data set 1 (80 stars), we derive empirical scaling relations incorporating asteroseismic parameters – the frequency of maximum oscillation power, large, and small frequency separations – alongside effective temperature and metallicity. To account for non-Gaussian uncertainties in the stellar parameters, we employ an ensemble machine learning approach that combines Ridge, Huber, and Gradient Boosting regressors to predict stellar mass, radius, and age, along with deriving the corresponding coefficients for the scaling relations. We assess coefficient uncertainties by combining jackknife and bootstrap resampling techniques. Using the resulting scaling relations, we predict the mass, radius, and age of 30 stars in data set 2. These predictions closely match observational data, even when the parameters slightly differ from those in the training set. The predicted stellar ages range from 2.29 to 10.09 Gyr, consistent with Solar-like oscillators. Our findings show that combining machine learning regression with resampling-based uncertainty estimation provides an effective and reliable approach for deriving scaling relations, offering valuable insights for stellar characterization in asteroseismic research.
- Open Access
Statistical Analysis of Magnetic Patches in Solar Active Regions Using the Complex Network Method
Identifying solar active regions (ARs), which consist of one or more pairs of magnetic patches with opposite polarities, is essential due to their significant role in dynamic solar atmospheric phenomena. In this study, we analyze ARs during their emergence and evolution on the solar surface using a complex network-based method known as Identifying Solar Magnetic Patches (ISMP). To examine the magnetic characteristics, we selected a subregion of 125 $\times$ 125 pixels centered on AR NOAA No., 1158, observed in 2011. Line-of-sight magnetogram data were obtained from the Helioseismic and Magnetic Imager (HMI) onboard the Solar Dynamics Observatory (SDO). Our statistical analysis reveals that the distributions of patch area, lifetime, and magnetic flux follow power-law behavior, with exponents approximately equal to $α$ = 2.14, 2.5, and 1.42, respectively. Furthermore, a Hurst exponent of 0.57 indicates the presence of long-range temporal correlations in the emergence of new magnetic patches.Open AccessStatistical Analysis of Magnetic Patches in Solar Active Regions Using the Complex Network Method
Identifying solar active regions (ARs), which consist of one or more pairs of magnetic patches with opposite polarities, is essential due to their significant role in dynamic solar atmospheric phenomena. In this study, we analyze ARs during their emergence and evolution on the solar surface using a complex network-based method known as Identifying Solar Magnetic Patches (ISMP). To examine the magnetic characteristics, we selected a subregion of 125 $\times$ 125 pixels centered on AR NOAA No., 1158, observed in 2011. Line-of-sight magnetogram data were obtained from the Helioseismic and Magnetic Imager (HMI) onboard the Solar Dynamics Observatory (SDO). Our statistical analysis reveals that the distributions of patch area, lifetime, and magnetic flux follow power-law behavior, with exponents approximately equal to $α$ = 2.14, 2.5, and 1.42, respectively. Furthermore, a Hurst exponent of 0.57 indicates the presence of long-range temporal correlations in the emergence of new magnetic patches. - Open Access
- 1 July 2025
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 279 (1) , 27
Solar Flare Prediction Using Long Short-term Memory (LSTM) and Decomposition-LSTM with Sliding Window Pattern Recognition
We investigate the use of long short-term memory (LSTM) and decomposition-LSTM (DLSTM) networks, combined with an ensemble algorithm, to predict solar flare occurrences using time series data from the GOES catalog. The data set spans from 2003 to 2023 and includes 151,071 flare events. Among approximately possible patterns, 7552 yearly pattern windows are identified, highlighting the challenge of long-term forecasting due to the Sun’s complex, self-organized-criticality-driven behavior. A sliding window technique is employed to detect temporal quasi-patterns in both irregular and regularized flare time series. Regularization reduces complexity, enhances large flare activity, and captures active days more effectively. To address class imbalance, resampling methods are applied. LSTM and DLSTM models are trained on sequences of peak fluxes and waiting times from irregular time series, while LSTM and DLSTM, integrated with an ensemble approach, are applied to sliding windows of regularized time series with a 3 hr interval. Performance metrics, particularly the true skill statistic (0.74), recall (0.95), and the area under the curve (AUC = 0.87) in the receiver operating characteristic, indicate that DLSTM with an ensemble approach on regularized time series outperforms other models, offering more accurate large-flare forecasts with fewer false errors compared to models trained on irregular time series. The superior performance of DLSTM is attributed to its ability to decompose time series into trend and seasonal components, effectively isolating random noise. This study underscores the potential of advanced machine learning techniques for solar flare prediction and highlights the importance of incorporating various solar cycle phases and resampling strategies to enhance forecasting reliability.
- 1 July 2025
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 279 (1) , 27
Open AccessSolar Flare Prediction Using Long Short-term Memory (LSTM) and Decomposition-LSTM with Sliding Window Pattern Recognition
We investigate the use of long short-term memory (LSTM) and decomposition-LSTM (DLSTM) networks, combined with an ensemble algorithm, to predict solar flare occurrences using time series data from the GOES catalog. The data set spans from 2003 to 2023 and includes 151,071 flare events. Among approximately possible patterns, 7552 yearly pattern windows are identified, highlighting the challenge of long-term forecasting due to the Sun’s complex, self-organized-criticality-driven behavior. A sliding window technique is employed to detect temporal quasi-patterns in both irregular and regularized flare time series. Regularization reduces complexity, enhances large flare activity, and captures active days more effectively. To address class imbalance, resampling methods are applied. LSTM and DLSTM models are trained on sequences of peak fluxes and waiting times from irregular time series, while LSTM and DLSTM, integrated with an ensemble approach, are applied to sliding windows of regularized time series with a 3 hr interval. Performance metrics, particularly the true skill statistic (0.74), recall (0.95), and the area under the curve (AUC = 0.87) in the receiver operating characteristic, indicate that DLSTM with an ensemble approach on regularized time series outperforms other models, offering more accurate large-flare forecasts with fewer false errors compared to models trained on irregular time series. The superior performance of DLSTM is attributed to its ability to decompose time series into trend and seasonal components, effectively isolating random noise. This study underscores the potential of advanced machine learning techniques for solar flare prediction and highlights the importance of incorporating various solar cycle phases and resampling strategies to enhance forecasting reliability.
- Open Access
- 19 February 2025
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 277 (1)
Galaxy Morphological Classification with Zernike Moments and Machine Learning Approaches
Classifying galaxies is an essential step for studying their structures and dynamics. Using GalaxyZoo2 (GZ2) fractions thresholds, we collect 545 and 11,735 samples in nongalaxy and galaxy classes, respectively. We compute the Zernike moments (ZMs) for GZ2 images, extracting unique and independent characteristics of galaxies. The uniqueness due to the orthogonality and completeness of Zernike polynomials, reconstruction of the original images with minimum errors, invariances (rotation, translation, and scaling), different block structures, and discriminant decision boundaries of ZMs’ probability density functions for different order numbers indicate the capability of ZMs in describing galaxy features. We classify the GZ2 samples, first into the galaxies and nongalaxies and second, galaxies into spiral, elliptical, and odd objects (e.g., ring, lens, disturbed, irregular, merger, and dust lane). The two models include the support vector machine (SVM) and 1D convolutional neural network (1D-CNN), which use ZMs, compared with the other three classification models of 2D-CNN, ResNet50, and VGG16 that apply the features from original images. We find the true skill statistic (TSS) greater than 0.86 for the SVM and 1D-CNN with ZMs for the oversampled galaxy–nongalaxy classifier. The SVM with ZMs model has a high-performance classification for galaxy and nongalaxy data sets. We show that the SVM with ZMs, 1D-CNN with ZMs, and VGG16 with vision transformer are high-performance (accuracy larger than 0.90 and TSS greater than 0.86) models for classifying the galaxies into spiral, elliptical, and odd objects. We conclude that these machine learning algorithms are helpful tools for classifying galaxy images. The Python notebooks are available on GitHub at https://github.com/hmddev1/machine_learning_for_morphological_galaxy_classification.
- 19 February 2025
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 277 (1)
Open AccessGalaxy Morphological Classification with Zernike Moments and Machine Learning Approaches
Classifying galaxies is an essential step for studying their structures and dynamics. Using GalaxyZoo2 (GZ2) fractions thresholds, we collect 545 and 11,735 samples in nongalaxy and galaxy classes, respectively. We compute the Zernike moments (ZMs) for GZ2 images, extracting unique and independent characteristics of galaxies. The uniqueness due to the orthogonality and completeness of Zernike polynomials, reconstruction of the original images with minimum errors, invariances (rotation, translation, and scaling), different block structures, and discriminant decision boundaries of ZMs’ probability density functions for different order numbers indicate the capability of ZMs in describing galaxy features. We classify the GZ2 samples, first into the galaxies and nongalaxies and second, galaxies into spiral, elliptical, and odd objects (e.g., ring, lens, disturbed, irregular, merger, and dust lane). The two models include the support vector machine (SVM) and 1D convolutional neural network (1D-CNN), which use ZMs, compared with the other three classification models of 2D-CNN, ResNet50, and VGG16 that apply the features from original images. We find the true skill statistic (TSS) greater than 0.86 for the SVM and 1D-CNN with ZMs for the oversampled galaxy–nongalaxy classifier. The SVM with ZMs model has a high-performance classification for galaxy and nongalaxy data sets. We show that the SVM with ZMs, 1D-CNN with ZMs, and VGG16 with vision transformer are high-performance (accuracy larger than 0.90 and TSS greater than 0.86) models for classifying the galaxies into spiral, elliptical, and odd objects. We conclude that these machine learning algorithms are helpful tools for classifying galaxy images. The Python notebooks are available on GitHub at https://github.com/hmddev1/machine_learning_for_morphological_galaxy_classification.
- Open Access
Galaxy Morphological Classification with Zernike Moments and Machine Learning Approaches
Classifying galaxies is an essential step for studying their structures and dynamics. Using GalaxyZoo2 (GZ2) fractions thresholds, we collect 545 and 11,735 samples in non-galaxy and galaxy classes, respectively. We compute the Zernike moments (ZMs) for GZ2 images, extracting unique and independent characteristics of galaxies. The uniqueness due to the orthogonality and completeness of Zernike polynomials, reconstruction of the original images with minimum errors, invariances (rotation, translation, and scaling), different block structures, and discriminant decision boundaries of ZMs' probability density functions for different order numbers indicate the capability of ZMs in describing galaxy features. We classify the GZ2 samples, firstly into the galaxies and non-galaxies and secondly, galaxies into spiral, elliptical, and odd objects (e.g., ring, lens, disturbed, irregular, merger, and dust lane). The two models include the support vector machine (SVM) and one-dimensional convolutional neural network (1D-CNN), which use ZMs, compared with the other three classification models of 2D-CNN, ResNet50, and VGG16 that apply the features from original images. We find the true skill statistic (TSS) greater than 0.86 for the SVM and 1D-CNN with ZMs for the oversampled galaxy-non-galaxy classifier. The SVM with ZMs model has a high-performance classification for galaxy and non-galaxy datasets. We show that the SVM with ZMs, 1D-CNN with ZMs, and VGG16 with vision transformer are high-performance (accuracy larger than 0.90 and TSS greater than 0.86) models for classifying the galaxies into spiral, elliptical, and odd objects. We conclude that these machine-learning algorithms are helpful tools for classifying galaxy images.
Open AccessGalaxy Morphological Classification with Zernike Moments and Machine Learning Approaches
Classifying galaxies is an essential step for studying their structures and dynamics. Using GalaxyZoo2 (GZ2) fractions thresholds, we collect 545 and 11,735 samples in non-galaxy and galaxy classes, respectively. We compute the Zernike moments (ZMs) for GZ2 images, extracting unique and independent characteristics of galaxies. The uniqueness due to the orthogonality and completeness of Zernike polynomials, reconstruction of the original images with minimum errors, invariances (rotation, translation, and scaling), different block structures, and discriminant decision boundaries of ZMs' probability density functions for different order numbers indicate the capability of ZMs in describing galaxy features. We classify the GZ2 samples, firstly into the galaxies and non-galaxies and secondly, galaxies into spiral, elliptical, and odd objects (e.g., ring, lens, disturbed, irregular, merger, and dust lane). The two models include the support vector machine (SVM) and one-dimensional convolutional neural network (1D-CNN), which use ZMs, compared with the other three classification models of 2D-CNN, ResNet50, and VGG16 that apply the features from original images. We find the true skill statistic (TSS) greater than 0.86 for the SVM and 1D-CNN with ZMs for the oversampled galaxy-non-galaxy classifier. The SVM with ZMs model has a high-performance classification for galaxy and non-galaxy datasets. We show that the SVM with ZMs, 1D-CNN with ZMs, and VGG16 with vision transformer are high-performance (accuracy larger than 0.90 and TSS greater than 0.86) models for classifying the galaxies into spiral, elliptical, and odd objects. We conclude that these machine-learning algorithms are helpful tools for classifying galaxy images.
- Open Access
Relationships of Earthquake Moment Magnitude with Body and Surface Waves Using MCMC Approach
Due to the saturation of the body ($m_b$) and surface ($M_S$) earthquake magnitudes, the moment magnitude ($M_W$) is a more convenient parameter for representing earthquake energies. We use the HRVD data, including 18,569 earthquakes recorded from 1976 January 1 to 2010 December 31, for five regions on the Earth (\textbf{Asia}-Oceania, Europe, Africa, North America, and South America). Applying the error propagation technique on the moment tensor, we estimate the standard error for $M_W$ showing a distribution that deviates from Gaussian errors which recommends using the Monte Carlo Markov chain (MCMC) to obtain model parameters instead of the least square approach. We investigate the linear relationship between the body and surface waves in small ($\le$ 6.1) and large ($>$6.1) magnitudes via MCMC. The slope varies from 0.811 to 1.565 ($m_b - M_W$) and 0.566 to 0.971 ($M_S - M_W$) across five regions. For 18,569 global earthquakes, we obtained the slope for $m_b \le 6.1$ ($m_b>$6.1) 0.863 (1.374), while we find the slope for $M_S\le$6.1 ($M_S >$ 6.1) 0.581 (0.921). These transition relationships of the magnitude moments are useful for increasing earthquake monitoring capacity.Open AccessRelationships of Earthquake Moment Magnitude with Body and Surface Waves Using MCMC Approach
Due to the saturation of the body ($m_b$) and surface ($M_S$) earthquake magnitudes, the moment magnitude ($M_W$) is a more convenient parameter for representing earthquake energies. We use the HRVD data, including 18,569 earthquakes recorded from 1976 January 1 to 2010 December 31, for five regions on the Earth (\textbf{Asia}-Oceania, Europe, Africa, North America, and South America). Applying the error propagation technique on the moment tensor, we estimate the standard error for $M_W$ showing a distribution that deviates from Gaussian errors which recommends using the Monte Carlo Markov chain (MCMC) to obtain model parameters instead of the least square approach. We investigate the linear relationship between the body and surface waves in small ($\le$ 6.1) and large ($>$6.1) magnitudes via MCMC. The slope varies from 0.811 to 1.565 ($m_b - M_W$) and 0.566 to 0.971 ($M_S - M_W$) across five regions. For 18,569 global earthquakes, we obtained the slope for $m_b \le 6.1$ ($m_b>$6.1) 0.863 (1.374), while we find the slope for $M_S\le$6.1 ($M_S >$ 6.1) 0.581 (0.921). These transition relationships of the magnitude moments are useful for increasing earthquake monitoring capacity.