STA 290 Seminar Series
Thursday, June 4, 2015, 4:10pm, MSB 1147 (Colloquium Room)
Refreshments at 3:30pm in MSB 4110 (Statistics Lounge)
Speaker: Alexandra Piryatinska
(Dept. Mathematics, San Francisco State University)
Title: Methodology for Segmentation of time series of arbitrary nature via $\epsilon$-complexity of continuous function its applications
Abstract: In this talk the problem of splitting time series of arbitrary nature (stochastic,
deterministic, or mixed) into segments generated by the same mechanism will be considered. The concept of $\epsilon$-complexity of continuous functions will be introduced. A characterization of this quantity for H\"{o}lder continuous functions will be given. On the basis of the $\epsilon$-complexity parameters, we propose a technique for the segmentation of time series that does not require any prior knowledge of how these series were generated. The results of simulations, and applications to EEG and stock market data, demonstrate the efficiency of the proposed methodology.
(Joint work with Boris Darkhovsky).