摘要
降水量是气象因素中极为重要的一项,降水量时间和空间上的不均匀分布在一定程度上将会导致一些地区的洪涝或干旱.基于函数型主成分分析的方法对月降水量数据进行应用分析.选取了全国31个城市23年的月降水量数据,通过引入傅里叶基函数,对其进行函数化得到降水量曲线,再利用函数型主成分分析研究.分析得到前四个主成分累积贡献为93.28%,可以解释原始数据大多数信息,降维效果比传统主成分分析更好.利用前四个函数型主成分以及对平均曲线的扰动情况,结合实际特征,给出部分地区降水量显著变化的时间.
Precipitation is an extremely important meteorological factor,the uneven distribution of precipitation over time and space will to some extent lead to flooding or drought in some areas.The method of functional principal component analysis is applied to analyze monthly precipitation data.The monthly precipitation data of 31 major cities in China for 23 years is selected,and the precipitation curve is obtained by introducing the Fourier basis function and functionalizing it.Then,functional principal component analysis was used for research.The analysis shows that the cumulative contribution of the first four principal components is 93.28%,which can explain most information of the original data,and the dimension reduction effect is better than the traditional principal component analysis.Finally,using the first four functional principal components and the perturbation of the average curve,combined with actual characteristics,the time of significant changes in precipitation in some regions is given.
作者
郑大钊
ZHENG Dazhao(School of Science,Qiqihar University,Qiqihar 161006,China)
出处
《高师理科学刊》
2023年第12期27-33,共7页
Journal of Science of Teachers＇College and University
基金
黑龙江省教育厅基本业务专项(135109228)。
关键词
函数型数据
月降水量
函数型主成分分析
functional data
monthly precipitation
functional principal component analysis
