基于多重比较检验的函数型数据方差分析  

Functional Analysis of Variance Based on Multiple Comparison Test

作　　者：白永昕 田茂再[2,3] Bai Yongxin;Tian Maozai(School of Statistics, Lanzhou University of Finance and Economics, Lanzhou 730020, China;a.Center for Applied Statistics, b School of Statistics, Renmin University of China, Beijing 100972, China)

机构地区：[1]兰州财经大学统计学院,兰州730020 [2]中国人民大学应用统计科学研究中心,北京100972 [3]中国人民大学统计学院,北京100972

出　　处：《统计与决策》2018年第10期62-65,共4页Statistics & Decision

基　　金：国家自然科学基金资助项目(11271368);国家社会科学基金重点项目(13AZD064);教育部高等学校博士学科点专项科研基金资助项目(20130004110007);北京市社会科学基金重大项目(15ZDA17);兰州商学院“飞天学者特聘计划”

摘　　要：随着现代社会可获得数据的大量性和连续性,函数型数据广泛地应用在各个领域。以甘肃省为研究对象,文章采用函数型数据方差分析对四个气候带的气温进行研究,并通过多重比较检验方法比较了各个气候带对气温的影响差异。结果表明：陇中地区全年气温受气候带的影响最小;河西走廊地区次之;陇南地区和甘南地区最大。由于受到气候带的影响,陇南地区和甘南全年气温较平均气温有较大的波动。气温在秋季和春季的波动较小,在夏季和冬季的波动相对较大。With the great availability and continuity of data in modern society, functional data is widely used in various fields. Taking Gansu Province as the research object, this paper uses the functional data analysis method of variance to research the temperature of four climatic zones, and compares the influence differences of each climatic zone on the temperature through multiple comparison test methods. The results show that the annual temperature in the middle of Gansu is the least affected by the climatic zone; the Hexi Corridor area is second; the southern areas of Gansu and Gannan areas are the largest; due to the influence of the climatic zone, the temperature of the southern Gansu and Gannan areas was more fluctuant than the average temperature in the whole year; temperatures fluctuate less in autumn and spring, and are relatively volatile in summer and winter.

关 键 词：函数型数据 方差分析 微分算子 多重比较检验 

分 类 号：O212.4[理学—概率论与数理统计]