【摘要】 目前时间序列数据取值大多是用一维数据,将一个时间段内的数据用一个数据值表示,这个值一般选取最高值、最低值、开始值、结束值、平均值等。一维数据表示的时间序列数据方便易用,分析简单,但同时也必然带来精度的下降。尤其在一些震动幅度大,震动频率高,对数据峰值又很敏感的领域,需要一种降低原始数据值误差的分段表示法。文章提出一种时间序列二维极值数据作为原始数据,通过数据的拓扑关系建立分段线性表示,实验证明这种方法能有效消除数据的极值误差,压缩率高,数据拟合度高,并能保持原始数据的趋势特征,满足数据动态增长需要。更多还原

【Abstract】 At present, all time series data values are mostly one-dimensional data, and the data in a period are represented by a data value. This value generally chooses the highest value, the lowest value, the beginning value, the end value, the average value and so on. The time series data represented by one-dimensional data is easy to use and simple to analyze, but at the same time, it will inevitably bring about a decline in accuracy. Especially in some areas where the vibration amplitude is large, the vibration frequency is high and the data peak value is sensitive, a piecewise representation is needed to reduce the error of the original data value. In this paper, a two-dimensional extreme data of time series is presented as the original data, and piecewise linear representation is established through the topological relationship of the data. The experiments show that this method can effectively eliminate the extreme error of the data and has high compression rate, high data fitting degree, and can retain the trend characteristics of the original data and meets the needs of dynamic growth of data.更多还原