摘要
离群点又称特异点、兴趣点、偏离点、新颖点、异常点等。通过离群点识别可发现异常事件与新现象。随着信息技术的发展和信息量爆炸式增长,通过识别数据中的离群点获得潜在信息成为研究热点。首先简要介绍几种主要的离群点识别方法,并分析各种方法的优缺点,为相关使用者学习、选择和改进算法提供参考。阐述离群点识别的研究热点和应用邻域,并分析现有算法在识别高维、空间和时序数据离群点的难点,便于研究者提出新的相关离群点识别方法。
Outliers are also called special points,interest points,deviations,novelty points,outliers,etc. Outlier identification can detect abnormal events and new phenomena. With the development of information technology and the explosive growth of information, potential information by identifying outliers in the data has become the research hotspot and it has attracted more and more attention. This paper briefly introduces several main outlier recognition methods,and concisely analyzes the advantages and disadvantages of each method,providing a reference for later users to learn,select and improve the algorithm. At the same time,the research hotspots and application neighborhoods of outlier recognition are described,and the difficulties of existing algorithms in identifying outliers in high-dimensional,spatial and temporal data are analyzed,which is convenient for relevant researchers to propose new outlier recogni-tion methods.
作者
黄强
叶青
聂斌
李欢
HUANG Qiang;YE Qing;NIE Bin;LI Huan(School of Computer Science,Jiangxi University of Traditional Chinese Medicine,Nanchang 330004,China)
出处
《软件导刊》
2019年第6期35-41,45,共8页
Software Guide
基金
国家自然科学基金项目(61562045)
江西省教育厅科学技术研究项目(160803)
江西省卫生计生委资助项目(2017A282)
江西中医药大学重点学科资助计划项目(2016jzzdxk015)
