节点文献
基于一类分类的时间序列异常值检测
Outliers Detection in Time Series Based on One-class Classification
【摘要】 论文结合相空间重构理论与一类分类方法提出一种时间序列中的异常值检测方法。该方法首先将时间序列映射到相空间,然后对相空间中的点实行一类分类,最后,根据KKT条件进行异常值检测。仿真实验结果表明了所给方法的可行性和有效性。
【Abstract】 A new method of outlier detection in time series is proposed in this paper,which is based on phase space reconstruction theory and one-class classification method.The method first maps time series to phase space, and then obtains a distinguish function by one-class classification in phase space, finally, outliers detection is proceed based on KKT conditions.The results of simulation experiments show the feasibility and effectiveness of the method.
【关键词】 异常值;
一类分类;
时间序列;
相空间;
【Key words】 outlier; one-class classification; time series; phase space;
【Key words】 outlier; one-class classification; time series; phase space;
【基金】 国家自然科学基金(编号:60075014);广东省自然科学基金(编号:021349)
- 【文献出处】 计算机工程与应用 ,Computer Engineering and Applications , 编辑部邮箱 ,2003年34期
- 【分类号】TP18
- 【被引频次】5
- 【下载频次】374