Constructed new metrics to measure re-identification risk arised in healthcare data and develop policies based on the risks. Developed a new metric to assess the security risk from health care records
Proposed a new framework using Branching process techniques for estimating the ancestral mean for Polymerase Chain Reaction data in random environments
Introduced a novel robust divergence based variational inference for social network model. Formulated Minimum Bregman Divergence Estimation, justified asymptotic theories, and conducted simulation studies and data analysis. Proposed Minimum Variational Divergence Estimator particularly for handling model mis-specification problems and implemented the methods to GLMM
Developed Hellinger distance based methods for streaming data from location scale family
Proposed new robust framework for models with latent structure and derived new EM-type algorithms. Conducted simulation study and data analysis using popular penalization techniques (e.g., lasso, scad, mcp) for high dimensional finite mixture regression model. Applied the new methods to horse treatment data for testing drug efficacy
Presentation titled ''Divergence Based Inference for High Dimensional GLMM''
Presentation titled ''Divergence Methods for Models with Latent Structure: Theory and Algorithms''
Presentation titled ''Privacy Analytics for Healthcare Data in Social Media via Divergence Techniques''
Presentation titled ''A New Framework for Re-identification Risk Estimation in Complex Healthcare Data''
Presentation titled ''Robust Estimate of Re-identication Risk In Complex Healthcare Data''
Job Titles Held: