R语言逐步回归报错：Error in is.data.frame(data) : object '.' not found

错误原因 #

你用dplyr的%>%管道把过滤后的数据直接传给lm()时，生成的线性模型对象里，data属性会被设置成管道的临时占位符.。这个.只在管道执行的瞬间存在，当后续调用step()函数时，临时对象已被回收，step()找不到数据对象就会抛出该错误。

解决方案 #

把过滤后的数据单独存成一个变量，再传入lm()，避免依赖管道的临时对象：

# 先保存过滤后的数据集
filtered_1999 <- country_tidy %>%
  filter(Year == "1999", 
         !is.na(`Life expectancy at birth, total (years)`))

# 基于保存好的数据集构建初始模型
lm_life_expectancy_1 <- lm(
  formula = `Life expectancy at birth, total (years)` ~ 1,
  data = filtered_1999
)

# 执行逐步回归
step(lm_life_expectancy_1, scope = ~ `Age at first marriage, female` + 
       `Age at first marriage, male` +
       `Birth rate, crude (per 1,000 people)` +
       `Births attended by skilled health staff (% of total)` +
       `Death rate, crude (per 1,000 people)` +
       `Fertility rate, total (births per woman)` +
       `GNI per capita, Atlas method (current US$)` +
       `Health expenditure per capita (current US$)` +
       `Health expenditure, total (% of GDP)` +
       `Hospital beds (per 1,000 people)` +
       `Incidence of HIV (% of uninfected population ages 15-49)` +
       `Incidence of tuberculosis (per 100,000 people)` +
       `Literacy rate, adult total (% of people ages 15 and above)` +
       `Mortality rate, adult, female (per 1,000 female adults)` +
       `Mortality rate, adult, male (per 1,000 male adults)` +
       `Mortality rate, infant (per 1,000 live births)` +
       `Physicians (per 1,000 people)` +
       `Population growth (annual %)` +
       `Population, female` +
       `Population, male` +
       `Population, total` +
       `Rural population` +
       `Rural population growth (annual %)` +
       `Urban population` +
       `Urban population growth (annual %)`)

关于缺失值的说明 #

数据中的缺失值不会直接导致这个错误。lm()默认会自动删除包含缺失值的行（行删除法），step()会基于lm()处理后的数据集继续运算。不过如果后续变量缺失值过多，可能会导致可用样本量骤降，但这和当前遇到的object '.' not found错误无关。

内容的提问来源于stack exchange，提问作者cheezeduckies