证券时间序列是证券交易价格的一组观测数据,是一种有其自身显著的特点的时间序列,针对这些特点我们提出一种基于分形理论与K线图形特点的分段方法,经过理论分析与实践证明其划分的证券时间序列分段有其合理性。在对时间序列数据压缩率很高的情况下,还能保持较好的拟合误差,并能较好地描述证券时间序列的走势特征。 Time sequence of securities is a set of observation data of the price of stock exchange, which is a kind of time series with its own salient features. In view of these characteristics, we propose a segmentation method based on fractal theory and K-line graph characteristics. After theoretical analysis and Practice has proved that its time-series segmentation of securities has its rationality. In the case of high compression rate of time series data, it also can maintain a good fitting error, and can better describe the trend characteristics of securities time series.