Self- and Super-organizing Maps in R: The kohonen Package

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

Authors:	Ron Wehrens, Lutgarde M. C. Buydens
Title:	[download] (22675)Self- and Super-organizing Maps in R: The kohonen Package
Reference:	Vol. 21, Issue 5, Oct 2007Submitted 2007-05-04, Accepted 2007-10-08
Type:	Article
Abstract:	In this age of ever-increasing data set sizes, especially in the natural sciences, visualisation becomes more and more important. Self-organizing maps have many features that make them attractive in this respect: they do not rely on distributional assumptions, can handle huge data sets with ease, and have shown their worth in a large number of applications. In this paper, we highlight the kohonen package for R, which implements self-organizing maps as well as some extensions for supervised pattern recognition and data fusion.
Paper:	[download] (22675)Self- and Super-organizing Maps in R: The kohonen Package (application/pdf, 1 MB)
Supplements:	[download] (1747)kohonen_2.0.0.tar.gz: R source package (application/x-gzip, 603.5 KB)
	[download] (1757)v21i05-figures.R: R code for producing the figures from the paper (application/zip, 1.1 KB)
	[download] (1843)v21i05.R: R example code from the paper (application/zip, 1.3 KB)
Resources:	BibTeX | OAI

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：Organizing package super maps Self especially attractive important features natural

1. library("kohonen")
   
2. data("wines")
   
3. wines.sc <- scale(wines)
   
4. 
   
5. set.seed(7)
   
6. wine.som <- som(data = wines.sc, grid = somgrid(5, 4, "hexagonal"))
   
7. plot(wine.som, main = "Wine data")
   
8. 
   
9. ####
   
10. readline("Press a key to continue")
    
11. data("nir")
    
12. attach(nir)
    
13. set.seed(13)
    
14. nir.xyf <- xyf(data = spectra,
    
15.                Y = composition[,2],
    
16.                xweight = 0.5,
    
17.                grid = somgrid(6, 6, "hexagonal"))
    
18. par(mfrow = c(1,2))
    
19. plot(nir.xyf, type = "counts", main = "NIR data: counts")
    
20. plot(nir.xyf, type = "quality", main = "NIR data: mapping quality")
    
21. 
    
22. ####
    
23. readline("Press a key to continue")
    
24. set.seed(13)
    
25. nir.xyf2 <- xyf(data = spectra,
    
26.                 Y = classvec2classmat(temperature),
    
27.                 xweight = .2, grid = somgrid(6, 6, "hexagonal"))
    
28. 
    
29. ####
    
30. readline("Press a key to continue")
    
31. water.predict <- predict(nir.xyf)$unit.prediction
    
32. temp.xyf <- predict(nir.xyf2)$unit.prediction
    
33. temp.predict <- as.numeric(classmat2classvec(temp.xyf))
    
34. 
    
35. par(mfrow = c(1,2))
    
36. plot(nir.xyf, type = "property", property = water.predict,
    
37.      main = "Prediction of water content", keepMargins = TRUE)
    
38. scatter <- matrix(rnorm(length(temperature)*2, sd=.1), ncol=2)
    
39. radii <- (temperature - 20)/250
    
40. symbols(nir.xyf$grid$pts[nir.xyf$unit.classif,] + scatter,
    
41.         circles = radii, inches = FALSE, add = TRUE)
    
42. 
    
43. plot(nir.xyf2, type = "property", property = temp.predict,
    
44.      palette.name = rainbow, main = "Prediction of temperatures")
    
45. scatter <- matrix(rnorm(nrow(composition)*2, sd=.1), ncol=2)
    
46. radii <- 0.05 + 0.4 * composition[,2]
    
47. symbols(nir.xyf2$grid$pts[nir.xyf2$unit.classif,] + scatter,
    
48.         circles = radii, inches = FALSE, add = TRUE)
    
49. 
    
50. ####
    
51. readline("Press a key to continue")
    
52. par(mfrow=c(1,2))
    
53. plot(nir.xyf2, "codes")
    
54. 
    
55. ####
    
56. readline("Press a key to continue")
    
57. data("yeast")
    
58. set.seed(7)
    
59. yeast.supersom <- supersom(yeast, somgrid(8, 8, "hexagonal"),
    
60.                            whatmap = 3:6)
    
61. classes <- levels(yeast$class)
    
62. colors <- c("yellow", "green", "blue", "red", "orange")
    
63. par(mfrow=c(3,2))
    
64. plot(yeast.supersom, type = "mapping",
    
65.      pch = 1, main = "All", keepMargins = TRUE)
    
66. for (i in seq(along = classes)) {
    
67.   X.class <- lapply(yeast,
    
68.                     function(x) subset(x, yeast$class == classes[i]))
    
69. 
    
70.   X.map <- map(yeast.supersom, X.class)
    
71.   plot(yeast.supersom, type = "mapping", classif = X.map,
    
72.        col = colors[i], pch = 1, main = classes[i], keepMargins = TRUE)
    
73. }
    
74. 
    
75. 
    
76. ####
    
77. readline("Press a key to continue")
    
78. graphics.off()
    
79. par(mfrow = c(1,3))
    
80. plot(wine.som, "changes", main = "Wine data: SOM")
    
81. plot(nir.xyf, "changes", main = "NIR data: XYF")
    
82. plot(yeast.supersom, "changes", main = "Yeast data: SUPERSOM")
    
83. 
    
84. 
    
85. ####
    
86. readline("Press a key to continue")
    
87. set.seed(7)
    
88. training.indices <- sample(800, 400)
    
89. training <- rep(FALSE, 800)
    
90. training[training.indices] <- TRUE
    
91. 
    
92. yeast.ssom1 <- supersom(lapply(yeast, function(x) subset(x, training)),
    
93.                         somgrid(4, 6, "hexagonal"),
    
94.                         whatmap = 3)
    
95. ssom1.predict <- predict(yeast.ssom1,
    
96.                          newdata = lapply(yeast,
    
97.                            function(x) subset(x, !training)),
    
98.                          trainY = subset(classvec2classmat(yeast[[7]]),
    
99.                            training))
    
100. prediction1 <- factor(classmat2classvec(ssom1.predict$prediction),
     
101.                       levels=levels(yeast$class))
     
102. confus1 <- table(subset(yeast$class, !training), prediction1)
     
103. confus1
     
104. sum(diag(confus1))
     
105. 
     
106. yeast.ssom2 <- supersom(lapply(yeast, function(x) subset(x, training)),
     
107.                         somgrid(4, 6, "hexagonal"),
     
108.                         whatmap = 3:6)
     
109. ssom2.predict <- predict(yeast.ssom2,
     
110.                          newdata = lapply(yeast,
     
111.                            function(x) subset(x, !training)),
     
112.                          trainY = subset(classvec2classmat(yeast[[7]]),
     
113.                            training))
     
114. prediction2 <- factor(classmat2classvec(ssom2.predict$prediction),
     
115.                       levels=levels(yeast$class))
     
116. confus2 <- table(subset(yeast$class, !training), prediction2)
     
117. confus2
     
118. sum(diag(confus2))

复制代码