Time Series Piecewise Linear Representation Based on Slope Extract Edge Point

In this paper, we introduce the slope which is widely used in analytic geometry, and bring forward a new Piecewise Linear Representation of Time Series which is based on Slope Extract Edge Point (SEEP).For the Time Series whose slope change range is relatively concentrate, SEEP representation has very good effects, compared with several other Piecewise Linear Representations, it’s fitting error with originality Time Series can be much smaller; for other Time Series, whose slope change range is relatively large, it’s fitting error with originality Time Series can not discrepant very much. This algorithm can be easy calculated and carried out, and it’s time complexity is only O(n).