我正在使用r中的step()函数。获取step()结束的变量列表应该保留在模型中的方法是什么?假设我有一个带有响应向量的数据框和几个预测变量X,Y和Z. Step()得出结论,只有X和Y应该包含在最终模型中。如何获得向量c('X','Y')
?
-
在回答第一个答案时,我尝试了$ coefficient和names()。两者都不起作用。
以下是print("names"); print(names(df_lm_2$coefficients))
[1] "(Intercept)" "How.often2" "How.often5"
[4] "How.often6" "How.often7" "How.often8"
[7] "Bathroom.s.2" "Bathroom.s.3" "Bathroom.s.5"
[10] "Bathroom.s.7" "Other.details..option.2.TRUE" "description_length"
[13] "has_descriptionTrue" "Other.details..option.6.TRUE" "Bedroom.s.2"
[16] "Bedroom.s.3" "Bedroom.s.4" "Bedroom.s.5"
[19] "Bedroom.s.6" "weekend1" "Other.details..option.7.TRUE"
[22] "Other.details..option.3.TRUE"
我最终想要的只是重要的列名,而不是每个因素的级别。对于上下文,以下是列名:
[1] "=------Colnames--------"
[1] "Bathroom.s." "Bedroom.s." "bid_price" "description_length"
[5] "has_description" "hour_of_day" "How.often" "log_bid_price"
[9] "month" "Other.details..option.1." "Other.details..option.2." "Other.details..option.3."
[13] "Other.details..option.4." "Other.details..option.5." "Other.details..option.6." "Other.details..option.7."
[17] "req_source" "Square.feet" "Type.of.home" "weekend"
答案 0 :(得分:0)
继续?step
的帮助页面中的示例,
lm1 <- lm(Fertility ~ ., data = swiss)
slm1 <- step(lm1)
你可以这样做:
names(slm1$coefficients)
请查看names(slm1)
以获得更多启示。