Premium
Detection of patterns in noisy time series
Author(s) -
Tingley M. A.,
Mclean L.
Publication year - 2001
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.2307/3316074
Subject(s) - series (stratigraphy) , context (archaeology) , white noise , noise (video) , binary number , statistics , independent and identically distributed random variables , algorithm , term (time) , mathematics , key (lock) , time series , computer science , amplitude , artificial intelligence , geography , random variable , arithmetic , physics , paleontology , archaeology , computer security , quantum mechanics , image (mathematics) , biology
In the context of a research project in ergonomy, myoelectric signals monitored over two to three hour periods gave rise to long noisy time series, which were smoothed using running medians. Tests developed by the authors show that the patterns displayed by the smoothed time series are not artifacts of smoothed white noise. Indeed, the smoothed series show amplitude fluctuations and short‐term correlations which are larger than those obtained by applying running medians to independent, identically distributed data. The key idea is that of reduction of data to binary signals.