灰色聚类及其R实现（人大经济论坛数据处理与分析中心实习作业）

非常清楚，谢谢楼主无私奉献。

huangqidong1987 发表于 2010-8-5 09:27
灰色聚类法是灰色系统一个重要应用分支之一。灰色系统最大的优势就在于其克服统计概率统计的弱点，从杂乱无 ...

学习一下

计算关联度部分的代码是否有瑕疵啊max(data)是否改为max(data[,j]，min(data)改为min(data[,j])?

1. ---
   
2. title: "灰色聚类及其R实现"
   
3. author: "屋檐滴语"
   
4. date: "2019/5/27"
   
5. documentclass: ctexart
   
6. output:
   
7.   rticles::ctex:
   
8.     fig_caption: yes
   
9.     number_sections: yes
   
10.     toc: yes
    
11. classoption: "hyperref,"
    
12. ---
    
13. 
    
14. # 载入数据
    
15. 
    
16. ```{r warning=FALSE, message=FALSE}
    
17. library(readxl)
    
18. data = read_xlsx("~/mydata/grey.xlsx", col_names = FALSE)
    
19. data = as.matrix(data)
    
20. rownames(data) = c("辽宁","山东","江苏","浙江","福建","广东")
    
21. colnames(data) = c("人均工业总产值","人均工业增加值","人均利润","人均所得税")
    
22. print(data)
    
23. ```
    
24. 
    
25. # 人均利润变正
    
26. 
    
27. ```{r}
    
28. data[,3] = data[,3] + 34
    
29. print(data)
    
30. ```
    
31. 
    
32. # 线性变换
    
33. 
    
34. ```{r}
    
35. for(j in 1:4){
    
36.   data[,j] = data[,j] / max(data[,j])
    
37. }
    
38. 
    
39. print(data)
    
40. ```
    
41. 
    
42. # 计算离差
    
43. 
    
44. ```{r}
    
45. for(j in 1:4){
    
46.   data[,j] =  max(data[,j]) - data[,j]
    
47. }
    
48. 
    
49. print(data)
    
50. ```
    
51. 
    
52. # 计算关联系数和关联度
    
53. 
    
54. ```{r}
    
55. data_gl = matrix(0, nrow = 6, ncol = 4)
    
56. 
    
57. for(i in 1:6)
    
58. for(j in 1:4)
    
59.   {
    
60.     data_gl[i,j] = ((max(data[,j]) - min(data[,j]))*0.5)/
    
61.       ((max(data[,j]) - min(data[,j])) * 0.5 + data[i,j] - min(data[,j]))
    
62.   }
    
63. 
    
64. rownames(data_gl) = c("辽宁","山东","江苏","浙江","福建","广东")
    
65. colnames(data_gl) = c("人均工业总产值","人均工业增加值","人均利润","人均所得税")
    
66. 
    
67. data_relation = matrix(0, nrow = 6, ncol = 1)
    
68. 
    
69. for(k in 1:6){
    
70.   data_relation[k,1] = sum(data_gl[k,])/4
    
71. }
    
72. 
    
73. colnames(data_relation) = c("关联度")  
    
74. data_cluster = cbind(data_gl, data_relation)
    
75. 
    
76. print(data_cluster)
    
77. ```
    
78. 
    
79. # 计算灰色聚类
    
80. 
    
81. ```{r}
    
82. data_grey = matrix(0, nrow=length(data_relation), ncol=length(data_relation))
    
83. data_sort = sort(as.numeric(data_relation), decreasing=T)
    
84. 
    
85. for(i in 1:6)
    
86. for(j in 1:6)
    
87. {
    
88. data_grey[i,j]<-(abs(data_sort[i]-data_sort[j])/data_sort[j])
    
89. }
    
90. 
    
91. data_grey_Ds<-matrix(0,nrow=length(data_relation),ncol=length(data_relation))
    
92. 
    
93. for(i in 1:6)
    
94. for(j in 1:6)
    
95. {
    
96. data_grey_Ds[i,j]<-data_grey[i,j]+data_grey[j,i]
    
97. }
    
98. 
    
99. data_grey_Rg<-matrix(0,nrow=length(data_relation),ncol=length(data_relation))
    
100. 
     
101. data_max<-max(data_grey_Ds)
     
102. 
     
103. for(i in 1:6)
     
104. for(j in 1:6)
     
105. {
     
106. data_grey_Rg[i,j]<-1-data_grey_Ds[i,j]/data_max
     
107. }
     
108. 
     
109. rownames(data_grey_Rg)<-c("广东","山东","辽宁","浙江","江苏","福建")
     
110. colnames(data_grey_Rg)<-c("广东","山东","辽宁","浙江","江苏","福建")
     
111. ```
     
112. 
     
113. # 画出系谱图
     
114. 
     
115. ```{r warning=FALSE, message=FALSE}
     
116. d<-as.dist(1-data_grey_Rg)
     
117. hc<-hclust(d,"single")
     
118. plot(hc,hang=-1)
     
119. ```

复制代码

收到，谢谢

收到 谢谢~~