关键词: 时间序列, 分段线性表示, 趋势转折点

Abstract: Piecewise Linear Representation（PLR） is an effective method to reduce the dimension of time series.A method of time series PLR algorithm is put forward based on the summarization of the characteristic of the series trends diversification.Firstly，this paper defines the trend transition points for the candidacy sets of time series’ segmenting points，and uses the distances between point and area to measure the weightiness of these points，then regards the important trend transition points that are chose according to specified threshold as the segmenting point for the PLR of time series.Compared with other six segmentation algorithms，the experiments show that the proposed algorithm appears better fitting quality and suitability.Especially for the time series which have the distinct transition point，the proposed algorithm has better ability of anti-noise-jamming.

Key words: time series, piecewise linearization representation, trend transition point