一种基于信息熵的时间序列分段线性表示方法

针对部分时间序列具有高维、大数据量及数据更新速度较快的特点, 导致在原始时间序列上难以进行数据挖掘的问题, 提出一种基于信息熵的时间序列分段线性表示方法——PLR_IE。该算法利用信息熵作为评判重要点数量的性能指标, 从序列中提取重要分段点的数量分布情况, 利用重要点组成的序列重新拟合原始时间序列, 为下一步数据挖掘提供基础。实验结果表明, 该方法能高效地提取出序列主要特征、拟合原始序列。

Piecewise linear representation based on information entropy of time series

For some time series with high dimension, large amount of data and data renewal speed characteristics, resulting in the original time series data mining on difficult problems, this paper presented a method piecewise linear representation of time series method based on information entropy, which was the piecewise linear representation of time series algorithm PLR_IE. The algorithm used the information entropy as an evaluation of important points of performance indicators, from sequence to extract important segment, the important point consisted of a sequence of fitting the original time series. Experiments show that this method can efficiently extract sequence, main characteristics, fitting the original sequence.