Motifs discovery for streaming time series

Abstract

The motif discovery approach is used to measure the correlation of the pair of consecutiveness in time series, which also aims to find all subsequences which are similar to the given one. However, alongwith the arrival of Industry 4.0 era, massive numbers of detectinginstruments in various fields are continuously producinga plenty number of time series streamingdata, the high dimensionality and continuousness of streamingtime series give rise to the potential threat for searchingeffectiveness. For thesereasons, wecomeupwithanovel motifs discovery approachfor streaming timeseries basedonpiecewiselinear representationwithturningpoints andskylineindex. As theexperimental results suggest, our approach is moreeffectivethan someother traditional methods.