一种自适应的高维离群点识别方法

doi:10.16088/j.issn.1001-6600.2020.02.012

摘要/Abstract

摘要： 针对传统的基于距离的离群点识别方法难以直接有效地应用于高维数据且识别效果受参数影响的问题,本文提出了一种自适应的高维离群点识别方法,该方法利用经遗传算法优化的高斯受限玻尔兹曼机将高维数据非线性地映射到低维空间,然后通过自适应的离群点识别方法在低维数据空间进行离群点识别。采用 UCI高维数据和中药高维数据进行验证实验,实验结果表明自适应的高维离群点识别方法能自适应地、有效地识别出高维数据中的离群点。

关键词: 离群点识别, 高斯受限玻尔兹曼机, 自适应算法, 遗传算法, 中医药信息

Abstract: Aiming at the problem that the traditional distance-based outlier recognition method can not be directly and effectively applied to high-dimensional data and the recognition effect is affected by parameters, an adaptive high-dimensional outlier recognition method is proposed, which uses genetic algorithm. The optimized Gaussian Restricted Boltzmann machine nonlinearly maps high-dimensional data to low-dimensional space, and then performs outlier recognition in low-dimensional data space by adaptive outlier recognition. UCI high-dimensional data and high-dimensional data of traditional Chinese medicine are used to verify the experiment. The experimental results show that the adaptive high-dimensional outlier recognition method can adaptively and effectively identify outliers in high-dimensional data.

Key words: outlier identification, GRBM, adaptive algorithm, genetic algorithm, Chinese medicine information

中图分类号:

TP301

叶青, 黄强, 聂斌, 李欢. 一种自适应的高维离群点识别方法[J]. 广西师范大学学报（自然科学版）, 2020, 38(2): 107-114.

YE Qing, HUANG Qiang, NIE Bin, LI Huan. An Adaptive High-Dimensional Outlier Recognition Method[J]. Journal of Guangxi Normal University(Natural Science Edition), 2020, 38(2): 107-114.

参考文献

[1] 张忠平,宋少英,宋晓辉.基于PCA及属性距离和的孤立点检测算法[J].计算机工程与应用,2009,45(17):139-141,243. DOI:10.3778/j.issn.1002-8331.2009.17.042.
[2] MEJIA A F,NEBEL M B,ELOYAN A,et al.PCA leverage: outlier detection for high-dimensional functional magnetic resonance imaging data[J].Biostatistics,2017,18(3):521-536.DOI:10.1093/biostatistics/kxw050.
[3] JOHNSTONE I M,PAUL D.PCA in high dimensions: an orientation[J].Proceedings of the IEEE,2018,106(8):1277-1292. DOI: 10.1109/JPROC.2018.2846730.
[4] JU Fujiao,SUN Yanfeng,GAO Junbin,et al.Image outlier detection and feature extraction via L1-norm based 2D probabilistic PCA[J].IEEE Transactions on Image Processing,2015,24(12):4834-4846.DOI:10.1109/TIP.2015.2469136.
[5] HUANG Haiping.Statistical mechanics of unsupervised feature learning in a restricted Boltzmann machine with binary synapses[J].Journal of Statistical Mechanics:Theory and Experiment,2017,2017(5):053302.DOI:10.1088/1742-5468/aa6ddc.
[6] 酆勇,熊庆宇,石为人,等.一种基于受限玻尔兹曼机的说话人特征提取算法[J].仪器仪表学报,2016,37(2):256-262.DOI: 10.19650/j.cnki.cjsi.2016.02.003.
[7] LI Ziqiang,CAI Xun,LIANG Ti.Gaussian-Bernoulli based convolutional restricted Boltzmann machine for images feature extraction[C]//Neural Information Processing:Lecture Notes in Computer Science vol 9948.Berlin:Springer,2016:593-602.DOI:10.1007/978-3-319-46672-9_66.
[8] 李敬微,顾晓辉,曹蕾,等.基于包络谱分析和高斯受限玻尔兹曼机的滚动轴承故障诊断方法[J].机械研究与应用,2016,29(2):87-90,93.DOI:10.16576/j.cnki.1007-4414.2016.02.029.
[9] 陈曦.基于高斯伯努利受限玻尔兹曼机的过程监测研究[D].杭州:浙江大学,2016.
[10]TRAN S N,BENETOS E,d′AVILA GARCEZ A.Learning motion-difference features using Gaussian restricted Boltzmann machines for efficient human action recognition[C]//2014 International Joint Conference on Neural Networks.Piscataway,NJ:IEEE Press,2014:2123-2129.DOI:10.1109/IJCNN.2014.6889945.
[11]LAMOUS-SWEENEY J D.Deep learning using genetic algorithms[D].Rochester,NY:Rochester Institute of Technology, 2012.
[12]TRAN L,FAN Liyue,SHAHABI C.Distance-based outlier detection in data streams[J].Proceedings of the VLDB Endowment,2016,9(12):1089-1100.DOI:10.14778/2994509.2994526.
[13]ZHANG Ying,ZHENG Hongyuan,DING Qiulin.Top-k distance-based outlier detection on uncertain data[C]//Cloud Computing and Security:Lecture Notes in Computer Science Vol 9483.Berlin:Springer,2015:521-535.DOI:10.1007/ 978-3-319-27051-7_45.
[14]YAN Qingli,CHEN Jianfeng,De STRYCKER L.An outlier detection method based on mahalanobis distance for source localization[J].Sensors,2018,18(7):2186.DOI:10.3390/s18072186.
[15]陆声链,林士敏.基于距离的孤立点检测研究[J].计算机工程与应用,2004,40(33):73-75,94.DOI:10.3321/j.issn: 1002-8331.2004.33.022.
[16]樊峰峰,李战怀,陈群,等.一种基于离群点检测的自动实体匹配方法[J].计算机学报,2017,40(10):2197-2211. DOI:10.11897/SP.J.1016.2017.02197.
[17]李春生,于澍,刘小刚.基于改进距离和的异常点检测算法研究[J].计算机技术与发展,2019,29(3):97-100. DOI:10.3969/j.issn.1673-629X.2019.03.021.
[18]张春霞,姬楠楠,王冠伟.受限波尔兹曼机[J].工程数学学报,2015,32(2):159-173.DOI:10.3969/j.issn.1005-3085.2015. 02.001.
[19]周志华.机器学习[M].北京:清华大学出版社, 2016.
[20]SUBRAMANIAM S,PALPANAS T,PAPADOPOULOS D,et al.Online outlier detection in sensor data using non-parametric models[C]//Proceedings of the 32nd International Conference on Very Large Data Bases.Seoul:VLDB Endowment,2006:187-198.
[21]ANNIE G.Anomaly detection based on machine learning:dimensionality reduction using PCA and classification using SVM[J].International Journal of Computer Applications,2012,47(21):5-8.DOI:10.5120/7470-0475.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed