Efficient Motif Discovery for Large-Scale Time Series in Healthcare
Abstract— Analyzing time series data can reveal the temporal behavior of the underlying mechanism producing the data. Time series motifs, which are similar subsequences or frequently occurring patterns, have signiﬁcant meanings for researchers especially in medical domain. With the fast growth of time series data, traditional methods for motif discovery are inefﬁcient and not applicable to large-scale data. This work proposes an efﬁcient Motif Discovery method for Large-scale time series < Final Year Projects 2016 > MDLats. By computing standard motifs, MDLats eliminates a majority of redundant computation in the related arts and reuses existing information to the maximum. All the motif types and subsequences are generated for subsequent analysis and classiﬁcation.