当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术23年3期

基于 AUC 的时间序列 Shapelets 提取方法
孙其法
(江苏联合职业技术学院徐州财经分院,江苏 徐州 221008)

摘  要:近年来,基于 Shapelets 的时间序列分类方法受到了广泛关注。现有的基于 Shapelets 转换的时间序列分类方法使用信息增益作为评价 Shapelets 的标准,并主要关注分类准确率,不能很好地适应不平衡时间序列分类。为了解决上述问题,使用 AUC 作为提取 Shapelets 的标准,并用于后续的数据集转换和分类,最后使用不平衡评价指标 F-measure 和 AUC 值对分类结果进行评价。结果显示,该方法能够很好地适应不平衡时间序列分类。


关键词:时间序列分类;Shapelets;不平衡数据分类;AUC



DOI:10.19850/j.cnki.2096-4706.2023.03.019


中图分类号:TP311                                         文献标识码:A                                    文章编号:2096-4706(2023)03-0083-04


AUC-based Time Series Shapelets Extraction Method

SUN Qifa

(Xuzhou Finance and Economics Branch of Jiangsu Union Technical Institute, Xuzhou 221008, China)

Abstract: In recent years, Shapelets-based time series classification methods have received extensive attention. Existing time series classification methods based on Shapelets transformation use information gain as a criterion for evaluating Shapelets, and mainly focus on classification accuracy, which cannot be well adapted to imbalanced time series classification. In order to solve the above problems, AUC (Area Under the Curve) is used as the criterion for extracting Shapelets and used for subsequent data set transformation and classification. Finally, the classification results are evaluated using the imbalanced evaluation index F-measure and AUC value. The results show that the method can adapt well to imbalanced time series classification.

Keywords: time series classification; Shapelets; imbalanced data classification; AUC 


参考文献:

[1] FAOUZI J. Time Series Classification:A review of Algorithms and Implementations [J/OL].Machine Learning(Emerging Trends and Applications),2022:1-34[2022-09-21].https://hal.inria.fr/hal- 03558165/document.

[2] 杨海民,潘志松,白玮 . 时间序列预测方法综述 [J]. 计算机科学,2019,46(1):21-28.

[3] 刘恒,侯越 . 贝叶斯神经网络在股票时间序列预测中的应用 [J]. 计算机工程与应用,2019,55(12):225-229+244.

[4] 潘晓,马昂,郭景峰,等 . 基于时间序列的轨迹数据相似性度量方法研究及应用综述 [J]. 燕山大学学报,2019,43(6):531-545.

[5] 付国庆 . 传感器网络时间序列数据计算及可视化研究 [D].大连:大连理工大学,2019.

[6] 杨梦晨,陈旭栋,蔡鹏,等 . 早期时间序列分类方法研究综述 [J]. 华东师范大学学报:自然科学版,2021(5):115-133.

[7] GRABOCKA J,SCHILLING N,WISTUBA M,et al. Learning time-series shapelets [C]//Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining.New York:Association for Computing Machinery,2014:392-401.

[8] HILLS J,LINES J,BARANAUSKAS E,e t a l . Classification of time series by shapelet transformation [J].Data Mining and Knowledge Discovery,2014,28(4):851-881.

[9] 金海波 . 一种时间序列 shapelets 快速发现算法 [J]. 太原科技大学学报,2019,40(3):229-235.

[10] YAN W H,LI G L,Wu Z D,et al. Extracting diverseshapelets for early classification on time series [J].World Wide Web, 2020,23(6):3055-3081.

[11] 闫汶和,李桂玲 . 基于 shapelet 的时间序列分类研究 [J].计算机科学,2019,46(1):29-35.

[12] 孙其法,闫秋艳,闫欣鸣 . 基于多样化 top-k shapelets 转换的时间序列分类方法 [J]. 计算机应用,2017,37(2):335-340.

[13] Hexagon.The UCR Time Series Classification Archive [EB/OL].[2022-08-25].https://www.cs.ucr.edu/~eamonn/time_series_ data_2018.


作者简介:孙其法(1991—),男,汉族,山东枣庄人,初级职称,硕士研究生,研究方向:数据挖掘、计算机网络。