Statistics and Data Science Seminar: "A General Cross Validation Framework with Applications to Dyadic CART and Trend Filtering"

Abstract: In this talk we will discuss a general cross validation framework for signal de-noising. We will then discuss how to apply this framework to two nonparametric regression methods Trend Filtering and Dyadic CART. The resulting cross validated versions would attain nearly the same rates of convergence as are known for the optimally tuned analogues. There did not exist any previous theoretical analyses of cross validated versions of Trend Filtering or Dyadic CART before this work. This framework is very general and potentially applicable to a wide range of estimation methods which use tuning parameters.

Hosts: Likai Chen and Debashis Mondal