Outlier detection algorithm based on pattern representation of time series
XU Rong-cong · Jisuanji yingyong yanjiu · 2007
This paper imported an algorithm which was based on the pattern representation of time series extract outlier value(PREOV).The pattern representation of time series itself had the function of compress data and keep the basic shape of time series,and it had a certain extent effect of noises removal.Based on PREOV could enormously increase the efficiency and veracity of algorithm,and got twice the result with half the effort.Besides,used some strategy of pruning,which made the algorithm's time complexity more lower.And this algorithm can be easy calculated and carry out,the training is needless and it can support the dynamic increase of time series.