将可选的权重参数传递给 model.frame

Question

我正在尝试编写一个函数，该函数采用可选的权重参数并将其传递给

model.frame

：

test <- function(ff, weights, data) {
  mm <- model.frame(ff, data, weights=weights)
  y <- model.response(mm)
  x <- mm[, 2]
  w <- mm[, 3:ncol(mm)-1]
  wt <- model.weights(mm)
  mat <- cbind(y, x, w, wt)
  return(sum(mat))
}

我收到一条错误消息

> test(ff, weight=wt, data=dtaw)
Error in model.frame.default(ff, data, weights = weights) : 
  invalid type (closure) for variable '(weights)'
Called from: model.frame.default(ff, data, weights = weights)

这是如何重现它

dtaw <- data.frame(y = runif(100))
dtaw$x <- runif(100)
dtaw$w1 <- runif(100)
dtaw$w2 <- runif(100)
dtaw$wt <- runif(100)

ff <- y ~ x + w1 + w2
test(ff, weight=wt, data=dtaw)

另外，有没有更好的方法从函数

model.frame

中的

mm

test

中提取第一个rhs变量和其余rhs变量？

Answer 1

这是更新的功能：

test <- function(ff, wt, data) {
  mm <- model.frame(ff, data, weights = wt)
  y <- model.response(mm)
  
  # Extract RHS variables
  rhs_vars <- all.vars(ff[[3]])
  x_var <- rhs_vars[1]
  w_vars <- rhs_vars[-1]
  
  x <- mm[[x_var]]
  w <- mm[w_vars]
  
  wt <- model.weights(mm)
  
  # Combine variables into a matrix
  mat <- cbind(y, x, w, wt)
  
  return(sum(mat))
}