关键词: 时间序列, 基于模型聚类, 马尔可夫模型, 动态时间弯折

Abstract: Compared to traditional classification problems，time series classification poses additional difficulties.A major difficulty is due to the fact that the time sequences are variable in length，making many traditional classification methods unable to apply directly.In this paper，the data conversion based on the cluster model is presented，and then model-based cluster analysis is given，time series modeling is maken with domain method，model parameters with composition of each vector is used to express sequence，the final time series matching algorithm analysis is given，classification algorithm is used for training and classification.Based on pipe flow of leakage point a time-series match，new methods is proposed，according continuity of similar samples time sequence is sorted，and the samples is inserted between adjacent samples.The new method is superior than the traditional method based on dynamic time bending.

Key words: time series, model based clustering, Markov model, dynamic time warping