How do I grab the variable names that are non-NA in a linear model?-CodePudding

library(datasets)
data(iris)
summary(iris)

iris$married = c(0)
iris$death = c(0)
iris$test = c(0)

regressor = lm(Sepal.Length ~ ., data = iris)
summary(regressor)

string = coef(summary(regressor))[2:summary(regressor)$fstatistic[2] 1,0]
string

names = c("Petal.Length","Petal.Width","Speciesversicolor","Speciesvirginica")

I can get it to output the non-na coefficients but I want to turn it into a list of characters like in 'names'. When I store it in 'string' it just becomes a weird num[1:4,0] type.

CodePudding user response：

You can use this code to extract the names of the coefficients from the summary:

library(datasets)
data(iris)
summary(iris)

iris$married = c(0)
iris$death = c(0)
iris$test = c(0)

regressor = lm(Sepal.Length ~ ., data = iris)
summary <- summary(regressor)

string = row.names(summary$coefficients)[c(-1)]
string

Output string:

[1] "Sepal.Width"       "Petal.Length"      "Petal.Width"       "Speciesversicolor" "Speciesvirginica"