使用多列的model.matrix

我想在model.matrix的data.frame中使用多列。使用多列的model.matrix

數據幀是這樣的：

df1 <- data.frame(id=seq(1,10,1), zip1=(round(runif(10)*100000,0)), zip2=(round(runif(10)*100000,0)) 
    ,var1=round(runif(10)*100,1),Sales=round(runif(10)* 10000,2)) 
df1$zip1 <- ifelse(nchar(as.character(df1$zip1)) < 5,paste("0",as.character(df1$zip1),sep=""),df1$zip1) 
df1$zip2 <- ifelse(nchar(as.character(df1$zip2)) < 5,paste("0",as.character(df1$zip2),sep=""),df1$zip2) 

df1$zip1 <- factor(df1$zip1) 
df1$zip2 <- factor(df1$zip2) 
dput(df1) 


> dput(df1) 
structure(list(id = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10), zip1 = structure(c(5L, 
1L, 8L, 3L, 7L, 9L, 2L, 6L, 10L, 4L), .Label = c("16667", "21922", 
"29100", "54398", "55447", "72607", "84667", "96562", "97012", 
"99125"), class = "factor"), zip2 = structure(c(7L, 4L, 2L, 6L, 
3L, 8L, 5L, 9L, 1L, 10L), .Label = c("0451", "0644", "14040", 
"17184", "18838", "42308", "48507", "50496", "64851", "89748" 
), class = "factor"), var1 = c(94.4, 39.6, 47.1, 74, 67.3, 53.4, 
28.7, 91.6, 47.1, 44.8), Sales = c(6394.03, 5575.65, 773.58, 
3181.43, 4992.56, 6627.01, 3313.13, 1585.1, 5080.98, 499.86)), .Names = c("id", 
"zip1", "zip2", "var1", "Sales"), row.names = c(NA, -10L), class = "data.frame")

，看起來像這樣：

> df1 
    id zip1 zip2 var1 Sales 
1 1 55447 48507 94.4 6394.03 
2 2 16667 17184 39.6 5575.65 
3 3 96562 0644 47.1 773.58 
4 4 29100 42308 74.0 3181.43 
5 5 84667 14040 67.3 4992.56 
6 6 97012 50496 53.4 6627.01 
7 7 21922 18838 28.7 3313.13 
8 8 72607 64851 91.6 1585.10 
9 9 99125 0451 47.1 5080.98 
10 10 54398 89748 44.8 499.86

我想這樣做：

df2 <- model.matrix(~df1[,-c(1,5)]) 
glmnet1 <- cv.glmnet(df2, df1[,"Sales"] 
    , family="gaussian", alpha=.95, nfolds=10, standardize = FALSE, type.measure="deviance")

不過，我可以」噸得到model.matrix採取多個變量：

> f1 <- formula(df1$Sales ~ df1[,c("zip1","zip2")]) 
> df2 <- model.matrix(f1) 
    Error in model.frame.default(object, data, xlev = xlev) : 
    invalid type (list) for variable 'df1[, c("zip1", "zip2")]'

有什麼建議嗎？

EDITS：

> f1 <- formula(Sales ~., data= df1[,-c(1,5)]) 
> df2 <- model.matrix(f1) 
Error in terms.formula(object) : '.' in formula and no 'data' argument

來源

2012-11-29 screechOwl

是這樣的：'model.matrix（Sales〜zip1 + zip2，data = df1）'你想要什麼？ – thelatemail

@thelatemail：這將起作用，我編輯了問題以更好地描述問題。我在〜100個變量上使用它，並且不想用var1 + var2 + ... + varx將它們寫出來。 – screechOwl

如果'model.matrix'接受「。」速記，然後修改'model.matrix（Sales〜。，data = df1）'（我認爲這是公式的正確語法，如果不是這樣的話）的建議。 –

可以指定多個變量的公式，如：

model.matrix(Sales ~ zip1 + zip2, data=df1)

如果你想節省打字的東西出來，嘗試：

formdf1 <- as.formula(paste("Sales ~ ", paste(names(df1)[2:3],collapse="+"))) 
formdf1 
#Sales ~ zip1 + zip2

，然後運行：

model.matrix(formdf1,data=df1)

來源

2012-11-29 20:25:02 thelatemail

require(glmnet) 
df2 <- df1[,-c(1)] # need the "Sales" column, drop it later 
# Could use df2 <- df1[ !names(df1) %in% exclude_vec ] 
glmnet1 <- cv.glmnet(model.matrix(Sales ~ . , df2)[, -1], df1[,"Sales"] , 
     family="gaussian", alpha=.95, nfolds=10, standardize = FALSE, 
     type.measure="deviance")

來源

2012-11-29 20:33:37

使用多列的model.matrix

回答

相關問題