基于时间序列的局部离群数据挖掘优化算法
Optimization Algorithm for Time Series-Based Local Outlier Data Mining Optimization Algorithm
姚红 1梁竹2
作者信息
- 1. 电子科技大学成都学院,四川 成都 611731
- 2. 重庆邮电大学计算机科学与技术学院,重庆 400065
- 折叠
摘要
针对数据量较大和数据维度较高导致离群数据挖掘困难的问题,提出基于时间序列的局部离群数据挖掘优化算法.将角度优化的全局嵌入算法和共同核主成分分析法相结合构建AOCKPCA降维算法,对海量高维时间序列降维处理;在蚁群算法中引入K-means算法,提升蚁群算法运算效率,降低不稳定性;将降维后的时间序列输入到优化后算法中,实现局部离群数据挖掘.实验结果表明,采用所提方法挖掘离群数据的准确率较高,误判的离群点个数较少,说明其挖掘效果较好.
Abstract
A local outlier data mining optimization algorithm based on time series is proposed to address the diffi-culty of outlier data mining caused by large data volumes and high data dimensions.Firstly,the global embedding al-gorithm based on angle optimization was combined with the common kernel principal component analysis method to form an AOCKPCA dimensionality reduction algorithm for reducing the dimensionality of massive high-dimensional time series.Then,the K-means algorithm was introduced to improve the computational efficiency of the ant colony al-gorithm and reduce the instability.Finally,the time series after dimension reduction was input into the optimized algo-rithm to realize the local outlier data mining.The experimental results show that the proposed method has high accura-cy in mining outlier data and a smaller number of misjudged outliers,indicating that its mining effect is better.
关键词
时间序列/局部离群数据挖掘/数据降维/蚁群算法Key words
Time series/Local outlier data mining/Data dimensionality reduction/Ant colony algorithm引用本文复制引用
基金项目
中国高校计算机教育MOOC联盟项目(B190205)
电子科技大学成都学院国腾创投基金(2021)(GTJG-04)
教育部产学合作协同育人项目(第二批)(2019)(2022015PC02470)
出版年
2024
NETL
NSTL
万方数据