在时间序列分段线性表示(PLR)基础上,提出一种新的基于特征点的分段方法,克服采用单一误差算法的模型失配问题,更加准确地反映过程状态的变化。
Based on PLR (piecewise linear representation)of time series, a novel important points-based method of segmentation is proposed. The problem of model mismatch is overcome ,and the obvious change of time series data is described precisely.