dr. P (Pariya) Behrouzi, PhD

I am an Assistant Professor in the Mathematical-Statistical Method group at Wageningen University (WU). My research focuses on developing statistical machine-learning frameworks to model complex processes and uncover underlying structures. Specifically, I work on probabilistic graphical models for graph structure learning, causal discovery (i.e., learning causal relationships from data), and causal representation learning from high-dimensional temporal data. By leveraging these techniques, I aim to enhance our understanding of complex systems and improve predictive modeling. I actively collaborate with researchers in plant sciences, food and nutrition, computational biology, and genetics to further refine and apply these methodologies in real-world contexts.

Previously, I was a postdoctoral researcher in the Mathematical-Statistical Method group at WU, where I worked on methodological advancements and software development for network inference in high-dimensional data. I received my PhD at the University of Groningen, focusing on extensions of graphical models for mixed discrete-and-continuous data and Granger causality in high-dimensional time-series data. During my PhD, I conducted research visits to the Mathematics Department at the University of Copenhagen and the MRC Biostatistics Unit at the University of Cambridge. Prior to that, I studied Mathematical Statistics in Iran.

I received the Hans van Houwelingen 2020 Award for the best Dutch biometry paper of 2018–2019, recognizing my work on copula graphical models for detecting epistatic selection, published in JRSS-C.

Software. I am the (co-)developer of several open-source R packages designed for statistical modeling and network analysis, including netgwas (Network-Based Genome-Wide Association Studies), netShiny (a tool for comparison and visualization of multiple networks), tsnetwork (Dynamic Chain Graph Models for Ordinal Time Series Data), and nutriNetwork (Structure Learning with Copula Graphical Models). These packages facilitate the analysis of complex data structures and have been used in various research domains.