为什么在gamm中找不到对象,但它在lm中工作?

时间:2017-01-09 14:31:21

标签: r function get gam mgcv

我正在努力编写一个使用get()

调用变量的R函数

假设我有这个数据框:

mydat = data.frame(y = rnorm(100), 
                   x = rnorm(100), 
                   day = sample(90:260, 100, replace = T), 
                   r1 = sample(seq(2008,2015,1), 100, replace = T),
                   r2 = sample(letters, 100, replace = T), 
                   r3 = sample(letters, 100, replace = T))

我想写一个返回gamm模型摘要的函数,比如这个

gamm_summary = function(data, response = "y"){
    require(mgcv)
    gamm_model = gamm(get(response) ~ s(day),
                      random = list(r1=~1, r2=~1, r3 =~1), data = data, method = "REML")

    summary(gamm_model$gam)
}


gamm_summary(mydat)

为什么这会给我错误:

  

get(响应)错误:找不到对象'响应'

但以下工作:

lm_summary = function(data, response = "y") {
    lm_model = lm(get(response) ~ x, data = data)
    summary(lm_model)
}

lm_summary(mydat)

问:为什么我的gamm函数无法工作?如何重写函数以使其有效?

1 个答案:

答案 0 :(得分:2)

require(mgcv)
gamm_summary = function(mydata, a1 = "y", b1 = "day", rnd1 = "r1", rnd2 = "r2", rnd3 = "r3"){
    df = data.frame(aa = mydata[[a1]], bb = mydata[[b1]], rnd1 = mydata[[rnd1]], rnd2 = mydata[[rnd2]], rnd3 = mydata[[rnd3]])
    gamm_model = gamm(aa ~ s(bb), data = df, random = list(rnd1 =~ 1, rnd2 =~ 1, rnd3 =~ 1), method = "REML")
    summary(gamm_model$gam)
}

gamm_summary(mydat, a1 = "x")