这是我写的一个ｒｅｇｒｅｓｓｉｏｎ　ｍｏｄｅｌ(mod4.lm)，下面两个是同一个模型的不同种写法，因为原模型中有自变量没有意义。　　
housing.value <- read.table("http://archive.ics.uci.edu/ml/machine-learning-databases/housing/housing.data", header=F)
colnames(housing.value) <- c("CRIM", "ZN", "INDUS", "CHAS", "NOX", "RM", "AGE", "DIS", "RAD", "TAX", "PTRAIO", "B", "LSTAT", "MEDV")
colnames(housing.value)

mod4.lm <- lm(MEDV~CRIM + ZN + CHAS + NOX + RM + DIS + PTRAIO + B + LSTAT, data=housing.value)
mod4.lm <- lm(MEDV~.-TAX-RAD-AGE-INDUS, data=housing.value)

我想把里面加入ｐａｉｒｗｉｓｅ　ｉｎｔｅｒａｃｔｉｏｎ　ｆａｃｔｏｒｓ　为了方便，重新定义了自变量。　但是，为什么我写的ｆｏｒ　ｌｏｏｐ　不对啊？

x1 <- housing.value$CRIM
x2 <- housing.value$ZN
x3 <- housing.value$CHAS
x4 <- housing.value$NOX
x5 <- housing.value$RM
x6 <- housing.value$DIS
x7 <- housing.value$PTRAIO
x8 <- housing.value$B
x9 <- housing.value$LSTAT

for(w in 1:9)
  for(v in 1:9)
{
   {
    a <- sum(x[w]*x[v])  #这是为了把那些成对的interaction factors 加和。
   }
}
   mod5.lm <- lm(MEDV~.-TAX-RAD-AGE-INDUS + a, data=housing.value)

特此悬赏解答，谢谢关注。