Research Area: Mathematical Statistics
My research program focuses on developing a better theoretical understanding of nonparametric statistical inference, using approximations when the sample size is large. A guiding principle is that complicated statistical models should be approximated by simple ones. This can be achieved within the theory of statistical experiments, using the concepts of local asymptotic normality and of Le Cam equivalence. Current efforts concentrate on models of dependent observations, such as stationary or locally stationary Gaussian sequences, with high or infinite dimensional parameter space. However there is also an interest in unsolved problems in the simple signal-plus-noise setting, such as sharp error asymptotics for adaptive nonparametric hypothesis testing.
Another line of research is quantum statistics, a field connected to recent progress in quantum computing and communication. Here the emphasis is on results for hypothesis testing and discrimination between quantum states, such as the quantum Chernoff bound.
[Homepage, Michael Nussbaum ]
Supported by National Science Foundation Grant No. DMS-1106460, Asymptotic Inference for Locally Stationary Processes
[Project summary | Description | References ]
[ Papers and manuscripts ]: includes publications under the grant and its predecessors since 06/2000.