Elvan Ceyhan Professor

Expertise
Dr. Elvan Ceyhan specializes in advanced statistical modeling and data science with a particular emphasis on graph-based machine learning and spatial statistical methods. His expertise includes the design, theoretical analysis, and implementation of network-driven algorithms for classification and clustering, addressing challenges such as class imbalance, structural overlap, and robustness to uncertainty. Dr. Ceyhan's foundational work in random geometric graphs, especially Proximity Catch Digraphs (PCDs), involves their theoretical characterization, domination properties, edge density, and connectivity.

He actively develops algorithmic frameworks for stochastic network optimization problems, including variants of the Canadian Traveler's Problem and Stochastic Obstacle Scene Problem, incorporating adversarial risks and path constraints. Additionally, his research applies nearest-neighbor and graph-theoretic methods to spatial pattern recognition, statistical clustering, and point-pattern analyses.

Dr. Ceyhan’s interdisciplinary contributions extend into biomedical data science, where he develops statistical methods for modeling and inference in high-dimensional medical and neuroimaging datasets, focusing on diagnostics, morphometric variability, and predictive analytics.

Research Interests

• Graph-Based Machine Learning: Algorithmic development for network-driven classification and clustering with emphasis on class imbalance, structural overlap, and robustness.
• Random Geometric Graphs: Theory and computational properties of Proximity Catch Digraphs (PCDs), including their domination number, edge and arc density, connectivity, and applications in pattern recognition.
• Network Optimization under Uncertainty: Stochastic and adversarial models for navigation and path planning, especially the Canadian Traveler’s Problem and Stochastic Obstacle Scene Problem.
• Spatial Statistics and Pattern Recognition: Graph-theoretic and nearest-neighbor methods for spatial clustering, spatial correlation analysis, and point-pattern recognition.
• Biomedical Data Science: Statistical inference and predictive modeling for high-dimensional neuroimaging and medical datasets.