Reza Arabi Belaghi

My research spans theoretical mathematical statistics, statistical methodology, and applied statistics, with a particular emphasis on developing statistically rigorous, computationally efficient, and interpretable methods for analysing complex and high-dimensional data. My work connects fundamental statistical theory with practical applications in medicine, epidemiology, agricultural sciences, and interdisciplinary research.

I have authored and co-authored more than 80 scientific publications, including research published in internationally recognized peer-reviewed journals. My research contributions encompass statistical estimation and inference, high-dimensional modelling, generalized linear models, shrinkage and penalized estimation, survival analysis, and machine learning, alongside applications in clinical prediction, cancer research, and population health.

Theoretical and Methodological Statistics
A central component of my research is the development and theoretical investigation of statistical estimation procedures, particularly in regression models involving correlated predictors, multicollinearity, high-dimensional covariates, and complex data structures.

My methodological research focuses on:

High-dimensional statistical modelling: Development and investigation of estimation and variable-selection methods for models with large numbers of predictors, including settings where the number of covariates is comparable to or exceeds the sample size.
Generalized linear models (GLMs): Theoretical and methodological developments in linear and nonlinear regression frameworks, including logistic, gamma, count-data, and zero-inflated regression models.
Shrinkage and penalized estimation: Development and evaluation of ridge-type, Liu-type, Stein-type, and preliminary-test estimation strategies to improve estimation accuracy and address multicollinearity and model uncertainty.
Statistical inference and asymptotic theory: Investigation of the theoretical properties of statistical estimators, including bias, variance, mean squared error, consistency, asymptotic distributions, and relative efficiency.
Variable selection and predictive modelling: Investigation of regularization, model selection, and post-selection estimation methods, with particular interest in identifying relevant predictors while improving estimation and predictive performance.
Statistical distributions and censored data: Estimation and inference for probability distributions under complex sampling and censoring schemes, using classical and Bayesian approaches.
An important objective of my theoretical research is to develop estimation procedures that perform reliably when conventional statistical methods encounter challenges arising from multicollinearity, high dimensionality, model uncertainty, or limited sample information.

Survival Analysis: Theory and Applications
Another major area of my research is survival analysis and time-to-event modelling, combining methodological development with applications in biomedical and clinical research.

My work encompasses Cox proportional hazards regression, parametric survival models, censored observations, regression-based risk estimation, and statistical approaches for analysing clinical outcomes over time.

I am particularly interested in developing and evaluating shrinkage and post-selection estimation procedures for high-dimensional survival models. This research addresses the challenges of estimating regression effects when many potential prognostic factors are present, including situations involving strong and weak predictor signals.

My methodological work investigates the statistical properties and practical performance of these estimators through theoretical analysis, simulation studies, and applications to real-world datasets.

The applied component of my survival research includes clinical prognosis, disease progression, treatment outcomes, hospitalization, and mortality, with particular emphasis on producing interpretable and statistically reliable estimates of risk.

Applied Statistics, Biostatistics, and Health Research
Alongside my theoretical research, I have extensive experience applying advanced statistical methods to important questions in medicine, epidemiology, and population health.

Cancer research and prediction: My research includes statistical modelling and machine learning applications for cancer diagnosis, biomarker identification, and risk prediction. I have contributed to studies involving prostate, ovarian, and gastric cancers, including the development and application of shrinkage estimators in cancer-related regression models and the use of machine learning for identifying potential molecular biomarkers.

Preterm birth prediction and maternal health: I have conducted population-based research on predicting preterm birth using traditional statistical models and machine learning methods. This includes developing and validating prediction models for nulliparous and multiparous women using large Canadian health datasets. These studies investigate maternal and pregnancy-related risk factors and compare the predictive performance of logistic regression, random forests, decision trees, and artificial neural networks.

Clinical and epidemiological research: My collaborative research includes longitudinal and time-to-event analyses in paediatric medicine, inflammatory bowel disease, and long-term care. I apply regression models, survival analysis, and advanced statistical methods to investigate disease progression, treatment effectiveness, clinical outcomes, and real-world evidence.

Machine learning and high-dimensional biomedical data: My applied research also explores statistical learning approaches for identifying important predictors, developing classification and prediction models, and evaluating model performance in complex biomedical datasets.

Agricultural, Environmental, and Interdisciplinary Applications
At SLU, I am interested in advancing the application of statistical methodology and machine learning in agricultural, environmental, animal, and natural sciences.

My expertise includes experimental design, regression analysis, longitudinal and hierarchical modelling, multivariate statistics, high-dimensional data analysis, and predictive modelling.

Through interdisciplinary research collaboration and statistical consulting, I support the development of scientifically appropriate analytical strategies, from study design and model selection to statistical inference, uncertainty quantification, and reproducible reporting.

Research Vision
My overarching research objective is to bridge mathematical statistical theory and real-world scientific applications. I am particularly interested in developing statistical methods that combine sound theoretical properties with strong practical performance, interpretability, and computational efficiency.

My research aims to advance statistical methodology while contributing to better scientific understanding, more reliable clinical prediction, and evidence-based decision-making across disciplines.

I also seek to strengthen the connections between classical statistical inference, high-dimensional modelling, survival analysis, and modern machine learning, ensuring that methodological innovation remains grounded in statistical theory and scientific reproducibility.

Teaching and pedagogical development are central components of my academic career. I have more than 12 years of university teaching experience across Sweden, Canada, and Iran, covering undergraduate, master's, doctoral, and professional education.

At SLU, I teach and coordinate courses in basic statistics, regression analysis, analysis of variance, and machine learning. My teaching emphasizes a strong understanding of statistical principles, practical data analysis, scientific interpretation, and the responsible use of statistical software.

Teaching areas

Basic statistics, probability, and statistical inference
Linear and multiple regression analysis
Analysis of variance and experimental design
Generalized linear models and advanced regression
Longitudinal data analysis and mixed-effects models
Survival analysis and biostatistics
Machine learning and predictive modelling
Statistical programming and reproducible analysis using R
Research methodology, study design, and sample size determination
Workshops and doctoral training

I actively develop and deliver workshops for PhD students, researchers, and academic staff through Statistics@SLU. These include:

Introduction to Machine Learning in Natural Sciences: Modelling and Applications (2023)
Advanced Regression Analysis in Natural Sciences Using R (2024)
Unsupervised Machine Learning with Applications in Agricultural Science Using R (2025)
Study Design and Sampling Techniques in Natural, Animal and Agricultural Sciences (2025)
Statistical Analysis Using Large Language Models: Guidelines and Best Practice (2026)
Teaching philosophy

My teaching philosophy is based on active learning, student engagement, critical thinking, and the connection between statistical theory and real-world research problems.

I aim to create an inclusive and supportive learning environment in which students develop not only technical statistical skills but also the ability to select appropriate methods, critically evaluate results, and communicate scientific findings.

I continuously develop my teaching through pedagogical training, student feedback, interdisciplinary collaboration, and reflection on teaching practices. I am particularly interested in how artificial intelligence and large language models can be integrated responsibly into statistical education while maintaining scientific rigour and independent critical thinking.