Hd=read.xlsx("支持向量机用数据.xlsx")#读取支持向量机用数据.xlsx
head(Hd)#查看数据

#归一化  
Hd=scale(Hd[,-1])

#查看变量之间的关系  
plot(Hd[,c("猪粮比价格变动率","玉米价格变动率（时差已调整）",  
           "存栏量变动率（时差已调整）",

cor(Hd[,c("猪粮比价格变动率","玉米价格变动率（时差已调整）",

n=nrow(Hd)
ntrain <- round(n*0.8) # 训练集
tindex <- sample(n,ntrain) # 筛选测试集样本

plot(Hd[,c("猪粮比价格变动率","玉米价格变动率（时差已调整）")] ,pch=ifelse

model <- svm(猪粮比价格变动率 ~ . , Hd)

mse <- function(error)  
{  
  sqrt(mean(error^2))
  
  
  predictionmse

## [1] 0.6789526

predictionmse=0  
jj=1  
for(i in seq(0,1,0.1)){  
  for(j in seq(0.1,1,0.1)){  
     
    model <- svm(Hd$"猪粮比价格变动率" ~ . 

which.min(predictionmse)

## [1] 10

,epsilon=1,cost=0.1)
points(Hd$"玉米价格变动率.时差已调整.", predictedY, col = "red", pch=4)

plot(Hd_predict[,c(3,2)] ,pch=ifelse(istrain==1,1,2))  
  
   
points(Hd_predict$"玉米价格变动率.时差已调整.", predictednew, col = "red", pch=4)