Fengnan Deng
Statistical Researcher & PhD Candidate
Statistical Researcher & PhD Candidate
PhD in Statistics
2019 - 2025 | Fairfax, VA
M.S. in Statistics
2017 - 2019 | New Brunswick, NJ
B.S. in Mathematics
2013 - 2017 | Jinan, China
Ensuring data privacy is crucial in modern statistical analysis, particularly when handling sensitive information. This project introduces a novel differential privacy framework that unifies multiple existing mechanisms while maintaining robust and efficient inference. By integrating privacy-preserving techniques with optimization methods, the work enhances the reliability of statistical learning under privacy constraints.
Credit risk assessment is essential in financial modeling, especially for large portfolios with diverse obligors. This project develops statistical models to evaluate total default risk using sparse high-dimensional factor structures. The research aims to provide a better theoretical understanding of portfolio risk under various economic conditions, particularly when the number of risk factors and obligor types grows.
Branching processes with immigration play a key role in modeling population dynamics, genetics, and epidemiology. This project explores the probability of rare events in such processes, particularly under different growth regimes. By deriving sharp large deviation estimates and studying probability-generating functions, the research uncovers new insights into the behavior of these stochastic models.
Hawkes processes are widely used for modeling self-exciting events, such as patient readmissions and disease outbreaks in healthcare. This project focuses on developing robust inference methods for Hawkes processes to improve their reliability in real-world applications. By incorporating density power divergence techniques, the research enhances model robustness against anomalies and data irregularities.
Deng, F. & Vidyashankar, A.N. (2024). Hellinger Differential Privacy: Private Estimation and Robust Inference. Journal Name. Preprint
Deng, F. (2023). Sharp Large Deviation Estimates for Branching Processes with Immigration. In Process
Computational and Financial Econometrics (CFE) & Computational and Methodological Statistics (CMStatistics) 2024
Topic: Sharp Large Deviations for Branching Processes with Immigration
Session: Branching and Related Processes Dec 14, 2024 King’s College London, UK
International Conference on Robust Statistics (ICORS) 2024
Topic: Hellinger Differential Privacy and Its Application to Hawkes Processes
Session: Differential Privacy and Robustness July 29, 2024 Virginia, USA
Western North American Region (WNAR) 2024
Privacy-Preserving Synthetic Hawkes Process Data
Session: Privacy Analytics: Theory and Applications June 10, 2024 Colorado, USA
Joint Statistical Meetings (JSM) 2024
Robust Privacy-Preserving Estimator for Hawkes Processes
Session: New Methods for Correlated Data Aug 6, 2024 Oregon, USA
2022 | Outstanding Graduate Student
2019-2021 | George Mason University
Courses: Probability and Statistics for Engineers and Scientists
July 2024 | George Mason University
Volunteer for the ICORS 2024 conference