Title: Elastic-band transform: A new way for multiscale method
Authors: Hee-Seok Oh - Seoul National University (Korea, South) [presenting]
Guebin Choi - Seoul National University (Korea, South)
Abstract: A new multiscale transformation method is presented for the statistical analysis of one-dimensional data such as time series and functional data under the concept of the scale-space approach. The proposed method uses regular observations (eye scanning) with a range of different intervals. The results, termed `elastic-band transform' can be considered as a collection of observations over various intervals (length of elastic-band) of viewing. It is motivated by a way that people look at an object such as a sequence of data repeatedly in order to overview a global structure of it as well as find some specific features of it. Some measures based on elastic-bands are considered for describing characteristics of data, and multiscale visualizations induced by the measures are developed for an understanding of data and detecting important structures of them. The proposed transform holds inherently some strengths for analyzing periodic signals because of its definition induced by a collection of regular observations; hence, statistical applications such as detection and signal extraction of periodic signals are studied.