本文關(guān)鍵詞： 函數(shù)型數(shù)據(jù) 導(dǎo)函數(shù) 區(qū)間 聚類分析　出處：《東北師范大學(xué)》2010年碩士論文　論文類型：學(xué)位論文

【摘要】： 實際生活中,越來越多領(lǐng)域中樣本的觀測數(shù)據(jù)具有函數(shù)型的特征,需要從函數(shù)視角進行分析.在此基礎(chǔ)上,與傳統(tǒng)分析方法相比較,函數(shù)型數(shù)據(jù)分析較少依賴假設(shè)條件,并且它的一個重要特征是可以求出導(dǎo)函數(shù)曲線,從中可以挖掘出更多的信息. 本文基于函數(shù)型數(shù)據(jù)的理論,總結(jié)了這類數(shù)據(jù)的基本特征和分析步驟,總結(jié)了函數(shù)型數(shù)據(jù)的聚類方法,分析了各自的優(yōu)缺點,并對基于軌跡模式的聚類方法進行了部分改進.最后將這些方法應(yīng)用于廣東省21個地級市的從業(yè)人口數(shù)據(jù)的聚類分析中,分析的結(jié)果對國家經(jīng)濟決策的制定具有一定的指導(dǎo)意義.
[Abstract]:In real life, the observed data in more and more fields have the characteristic of functional type, which need to be analyzed from the perspective of function. On this basis, compared with the traditional analysis method, the analysis of functional data is less dependent on hypothetical conditions. And one of its important features is that the derivative curve can be obtained and more information can be mined from it. Based on the theory of functional data, this paper summarizes the basic characteristics and analysis steps of this kind of data, summarizes the clustering method of functional data, and analyzes their advantages and disadvantages. Finally, these methods are applied to the cluster analysis of working population data of 21 cities in Guangdong Province. The results of the analysis have certain guiding significance for the formulation of national economic decision-making.
【學(xué)位授予單位】：東北師范大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2010
【分類號】：F224;C924.2

【參考文獻】

相關(guān)期刊論文 前1條

1 嚴(yán)明義;;函數(shù)性數(shù)據(jù)的統(tǒng)計分析:思想、方法和應(yīng)用[J];統(tǒng)計研究;2007年02期

，

本文編號：1554483