一種加權聚類分析的應用分析

信息計算科學論文編號:XXLW050 論文字數:8551,頁數:28

目  錄
摘     要 i
Abstract ii
目  錄 iii
第一章  緒論 1
1.1  引言 1
1.2  基本原理 2
1.2.1 關于聚類分析 2
1.2.2 統計軟件 3
第二章  加權聚類分析 4
2.1  數據收集 4
2.2  加權聚類分析 5
2.2.1 聚類分析 5
2.2.2 結合變異系數的加權聚類分析 6
2.2.3 結論 7
第三章  加權聚類分析的應用 8
3.1  杭州航空物流的發展及研究意義 8
3.2  數據收集 8
3.3  數據分析 11
3.4  加權聚類結果分析 14
第四章  結論 15
致  謝 16
參考文獻 17

摘     要
 聚類分析是研究“物以類聚”的一種科學有效的方法。人類在認識世界的過程中，需要把某些方法相似的東西歸成類，以便從中發現規律性，進而達到認識和改造世界的目的。早先，人們分類主要是憑借經驗，做些定性分析，很少利用數據做定量的研究，隨著科學發展和計算機的普及，利用數學方法研究分類問題成為可能，聚類分析作為多元分析的一個重要分支，發展迅速。以前研究人員對總體數據進行聚類分析時，主觀的給定權重，這種聚類效果往往不是很好。因此，本文提出了一種聚類的新方法，采用變量的變異系數作為權重以提高聚類的效果，并通過實例證明這種方法確實有效。本文將該方法運用到幫助企業在運輸包裹中選擇更準確的路線中，通過SPSS統計軟件進行聚類分析，得出相應的結論。

關鍵詞：聚類分析  權重  變異系數
Abstract
 Cluster analysis is a kind of scientific and effective method to study “things of one kind come together”. Whenever Human beings are during the process of recognizing the world, they need to make some similar things into one category, in order to discover the rules, and then to realize and reconstruct the world on purpose. Previously, people classified things mainly by virtue of experience, and did some qualitative analysis, but rarely made use of data to do the quantitative research. With the development of science and the popularization of computer, utilizing mathematical methods to solve classification problems become possible, cluster analysis, as an important branch of the multi-analysis, is developing rapidly. When the research on clustering analysis before were subjectively giving weight, this clustering effect is always not very good. Therefore, suggesting a new method of calculating the weight, using the coefficient of variation, and through the real examples to prove that this method is effective. Thus, applying the new method calculating of the weight to help company in the transport of choice parcels more accurate line, through the SPSS statistical software to do some cluster analysis, it can draw to the  corresponding conclusions.

Keywords：Cluster analysis; Weight Coefficient; Coefficient of variation