A common way to represent a time series is to divide it into shortduration blocks, each of which is then represented by a set of basis functions. A limitation of this approach, ho...
Monitoring and correlation of streaming data from multiple sources is becoming increasingly important in many application areas. Example applications include automated commodities...
We analyze the Synchrosqueezing transform, a consistent and invertible time-frequency analysis tool that can identify and extract oscillating components (of time-varying frequency...
Eugene Brevdo, Neven S. Fuckar, Gaurav Thakur, Hau...
Segmentation of CSF and pulsative blood flow, based on a single phase contrast MRA (PC-MRA) image can lead to imperfect classifications. In this paper, we present a novel automated...
Ali Gooya, Hongen Liao, Kiyoshi Matsumiya, Ken Mas...
In this paper we propose a scaling-up method that is applicable to essentially any induction algorithm based on discrete search. The result of applying the method to an algorithm ...