大气和海洋科学快报（英文版）2023，Vol.16Issue(4)：1-7,7.DOI:10.1016/j.aosl.2022.100292
A machine learning approach to quality-control Argo temperature data
A machine learning approach to quality-control Argo temperature data
摘要
Abstract
本文提出了一种基于机器学习的Argo浮标温度异常值检测方法.该方法采用机器学习无监督算法高斯混合模型对Argo浮标数据进行聚类分析,并构建包围所有数据点的最小多边形的凸包.基于射线投影算法实现点在多边形内分析,通过自动识别数据点位于凸包内外来判断该数据点数据质量的好坏.本文采用南海区域Argo浮标数据对该方法进行测试,结果表明该方法可以识别70％以上的包含异常值的温度剖面,同时自动标记出各异常值点.关键词
质量控制/机器学习/异常值检测/高斯混合模型/凸包/点在多边形内Key words
Quality control/Machine learning/Outlier detection/Gaussian mixture model/Convex hulls/Point-in-polygon引用本文复制引用
Qi Zhang,Chenyan Qian,Changming Dong..A machine learning approach to quality-control Argo temperature data[J].大气和海洋科学快报（英文版）,2023,16(4):1-7,7.基金项目
This research was funded by a project supported by the Southern Ma-rine Science and Engineering Guangdong Laboratory(Zhuhai)[grant number SML2020SP007],the National Natural Science Foundation of China[grant number 41906167],and the Startup Foundation for Intro-ducing Talent of Nanjing University of Information Science and Tech-nology[grant number 2018r077]. （Zhuhai）
