金融数据时间序列相似性度量的应用研究 |
| |
引用本文: | 肖娜,都泳涛. 金融数据时间序列相似性度量的应用研究[J]. 数字社区&智能家居, 2013, 0(9): 5600-5604 |
| |
作者姓名: | 肖娜 都泳涛 |
| |
作者单位: | 同济大学CAD研究中心,上海201804 |
| |
摘 要: | 从应用角度对时间序列数据挖掘中的关键技术一相似性度量一进行了研究。实现了对时间序列的分段线性表示,并将其用于当前主要的几种时间序列距离度量算法。通过将各距离度量算法用于股票收盘数据分析实验,得出实验数据。通过对实验结果的分析并结合各算法的原理,对各方法的适用情况和执行效率进行了分析及比较。通过分析可知,每种算法有自己的特点及适用情况。对于实际应用,应根据实际需求选择合适的距离度量算法。
|
关 键 词: | 时间序列 数据挖掘 分段线性表示 相似性度量 编辑距离 |
A Similarity Measure Research for Financial Time Series |
| |
Affiliation: | XIAO Na, HAO Yong-tao (CAD Research Center of Tongji University, Shanghai 201804, China) |
| |
Abstract: | Similarity measurement is the key technology in time series data mining. Present most existing time-sequence distance measurement algorithms, in this paper, are implemented using piecewise linear representation of financial data time series. Through the analysis of experimental results and according to the principle of each algorithm, applicable situation and the execu- tion efficiency of each method are analyzed and compared. The analysis shows that each algorithm has its own applicable condi- tion. For practical application, the distance measurement algorithm should be chosen according to actual needs. |
| |
Keywords: | time series data mining si~nilarity piecewise linear representation edit distance |
本文献已被 维普 等数据库收录! |