Student Seminar Series
DATE: Wednesday November 30th, 2016, 12:00pm
LOCATION: MSB 1147 (Colloquium Room).
SPEAKERS: Ozan Sonmez, PhD Candidate, Statistics, UC Davis
TITLE: “Detecting and dating structural breaks in functional data without dimension reduction”
ABSTRACT: We propose methodology for conducting change point analysis with functional data that is “fully functional” in the sense that it does not rely on dimension reduction techniques. A thorough asymptotic theory is developed for a fully functional change point test statistic, as well as for a break date estimator assuming a fixed break size and a shrinking break size. This final result is used to derive a confidence interval for the unknown break date. The main results highlight that the fully functional procedures perform best under conditions when analogous fPCA based estimators are at their worst, namely when the feature of interest is orthogonal to the leading principal components of the data. The theoretical findings are confirmed by means of a Monte Carlo simulation study in finite samples. An application to one-minute resolution intra- day cumulative log-returns of Microsoft stock data illustrates the practical relevance of the proposed fully functional procedure.
This seminar series is organized by PhD Students Irene Kim and Clark Fitzgerald.