2014 Statistics Ph.D. recipient
Carnegie Mellon University
I worked on smooth generalizations of post-stratication for capture-recapture applications. Find my main working paper here.
My research is potentially relevant for the design and evaluation of the U.S. Decennial Census. In particular, one can estimate how many people are missed in the census by comparing census records against administrative files such as IRS tax returns. CMU's "Census node" is my immediate family, academically speaking.
My approach to capture-recapture has required me to become interested in log-linear model selection and fast multinomial local logistic regression.