Statistical Learning Problems with Complex Stochastic Models

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

Big data is having a profound impact on scientific research and knowledge discovery. And while big data poses many statistical and computational challenges, it also presents unprecedented opportunities for statistics and data science. The investigator will focus on emerging scientific problems through the development of novel statistical and computational means, and address the challenges that arise in solving data intensive complex problems. The research in this project on finance statistics and computational algorithms is motivated by solving practical problems, and will yield cutting-edge statistical techniques and effective computational tools. The investigator actively participates in activities to integrate research with student training and applies the research outcomes to fields like finance and deep learning. The investigator will conduct novel research on stochastic gradient descent algorithms and unified models for financial data. The research goals are to develop innovative statistical methodologies, computing techniques, and theories for: 1) unified stochastic models for combined inference based on both high-frequency and low-frequency financial data, and 2) statistical and computational analysis of stochastic gradient descent algorithms with applications to machine learning in particular deep learning. The investigator intends to establish theoretically-supported statistical methodologies and computational procedures, and significantly advance computational and statistical understanding to the proposed research problems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.