-
Notifications
You must be signed in to change notification settings - Fork 0
/
LGandSVM.R
138 lines (108 loc) · 4.11 KB
/
LGandSVM.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
# SVM
if(!suppressWarnings(require('kernlab')))
{
install.packages('kernlab')
require('kernlab')
}
n <- 1000 # 样本数
p <- 2 # 变量数
sigma <- 1 # 分布的标准差
meanpos <- 0 # 正样本分布的均值
meanneg <- 3 # 负样本分布的均值
npos <- round(n/2) # 正样本数目
nneg <- n-npos # 负样本数目
# 生成正样本
##设置种子，为了试验可重复性
set.seed(1234)
xpos <- matrix(rnorm(npos*p,mean=meanpos,sd=sigma),npos,p)
##生成负样本
set.seed(1234)
xneg <- matrix(rnorm(nneg*p,mean=meanneg,sd=sigma),npos,p)
##正样本和负样本合并
x <- rbind(xpos,xneg)
# 生成标签
y <- matrix(c(rep(1,npos),rep(-1,nneg)))
# Visualize the data
plot(x,col=ifelse(y>0,1,2))
ntrain <- round(n*0.8) # number of training examples
tindex <- sample(n,ntrain) # indices of training samples
xtrain <- x[tindex,]
xtest <- x[-tindex,]
ytrain <- y[tindex]
ytest <- y[-tindex]
library(kernlab)
svp <- ksvm(xtrain,ytrain,type="C-svc",kernel='vanilladot',C=100,scaled=c())
plot(svp,data=xtrain)
ypred <- predict(svp,xtest) ##应用到测试集
table(ytest,ypred) ##利用table()生成混淆矩阵
sum(ypred==ytest)/length(ytest) ##模型预测的准确率
if(!suppressWarnings(require('ROCR')))
{
install.packages('ROCR')
require('ROCR')
}
ypredscore <- predict(svp,xtest,type="decision")
pred <- prediction(ypredscore,ytest)
perf <- performance(pred, measure = "tpr", x.measure = "fpr")
plot(perf)
---------------------------------------------------
if(!suppressWarnings(require(mlbench)))
{
install.packages('mlbench')
require(mlbench)
}
# library("e1071")
# data(Glass, package="mlbench")
Glass= read.table("tax_8OTU0.1.txt.T.Group.svf",header = T,check.names=F,as.is=F)
Glass$Group = as.factor(Glass$Group)
##第二步：数据集划分：训练集和测试集
# colnames(train_svm) <- c(paste('p',c(1:(nday*96+ex))),'f')
index <- 1:nrow(Glass)
testindex <- sample(index, trunc(length(index)/3))
testset <- Glass[testindex,]
trainset <- Glass[-testindex,]
##第三步：构建SVM模型
library("e1071")
svm(Group ~ ., data = trainset)
svm.model <- svm(Group ~ ., data = trainset, cost = 1000, gamma = 1, kernel = "polynomial" )
##第四步：SVM模型应用到测试数据集
svm.pred <- predict(svm.model, testset)
##第五步：模型结果评估
##1混淆矩阵
table(pred = svm.pred,true= testset[,10])
##2计算Accuracy和Kappa值
classAgreement(table(pred = svm.pred,true= testset[,10]))
###绘制ROC曲线
library("pROC")
svm_roc <- roc(testset$Group,as.numeric(svm.pred))
plot(svm_roc, print.auc=TRUE, grid=c(0.2, 0.2),grid.col=c("green", "red"),auc.polygon.col="skyblue", print.thres=TRUE,main='SVM模型ROC曲线 kernel = polynomial',xlim=c(1,0),ylim=c(0,1) ,add=T)
#####
#逻辑回归3
lg=glm(Group ~ ., data = trainset,family = "binomial")
lg.pred <- predict(lg, testset)
table(pred = lg.pred,true= testset[,10])
lg_roc <- roc(testset$Group,as.numeric(lg.pred))
plot(lg_roc, add=T,auc.polygon=TRUE,print.auc=TRUE,auc.polygon.col="skyblue", print.thres=TRUE,print.auc.x=0.3,print.auc.y=0.3,asp = NA) #RF
######
#####
#随机森林1
plot(rocc,print.auc=TRUE,col="red", print.thres=TRUE,print.auc.x=0.86,print.auc.y=0.86,auc.polygon.col="skyblue",asp = NA,main='ROC曲线', bty = "n" ) #RF
#####
plot_colors <- c("blue","red","black")
legend(x = "bottom",inset = 0,
legend = c("SVM:72.9%", "Logistic: 71.0%", "RandomForest:80.0%"),
col=plot_colors, lwd=7, cex=.7, horiz = TRUE)
# －－－－－－－－－－－－－－－－－－－－－－－－－－
#测试集
#pre_svm <- predict(wine_svm,newdata = test_data)
pre_svm <- predict(svm.model, testset)
#obs_p_svm = data.frame(prob=pre_svm,obs=test_data$等级)
obs_p_svm = data.frame(prob=pre_svm,obs=testset)
obs_p_svm = data.frame(prob=pre_svm,obs=testset$等级)
###输出混淆矩阵
#table(test_data$等级,pre_svm,dnn=c("真实值","预测值"))
table(testset$等级,pre_svm,dnn=c("真实值","预测值"))
###绘制ROC曲线
#SVM2
svm_roc <- roc(testset$等级,as.numeric(pre_svm))
plot(svm_roc, add=T,print.auc=TRUE, auc.polygon=TRUE, grid=c(0.1, 0.2),auc.polygon.col="gray",col="blue",, print.thres=TRUE,print.auc.x=0.9,print.auc.y=0.7,)