CODEWITHSUNDEEP
rlinear-regression
Justin Fletcher
Justin Fletcher

Reputation: 2409

Why won't lda() accept a string as it's 'formula' argument?

I'm doing some regression analysis and I've come across some strange behavior from the lda function in the MASS library. Specifically, it seems to be unable to accept a string as it's formula argument. This doesn't appear to be a problem for the base glm functions. I've constructed a small example using iris to illustrate the point.

library(MASS)
myForm<-"Species~Petal.Length"
# Disregard the warnings from this line, they're an artifact of the example. It works.
lgrIris<-glm(formula=myForm, data=iris, family="binomial")
# Breaks. 
ldaIris<-lda(formula=myForm, data=iris)

The final line above throws:

Error in lda.default(formula = myForm, data = iris) : 
argument "x" is missing, with no default

Which, judging from the documentation, seems to indicate that lda doesn't think it's been provided a formula argument. Does anyone know why this is happening, or how to fix it?

Upvotes: 1

Views: 310

Answers (2)

user3710546
user3710546

Reputation:

You can turn "myForm" into a formula using as.formula():

myForm <- "Species~Petal.Length"
class(myForm)
# [1] "character"
myForm <- as.formula(myForm)
class(myForm)
# [1] "formula"
myForm
# Species ~ Petal.Length

lda(formula=myForm, data=iris)
# Call:
# lda(myForm, data = iris)

# Prior probabilities of groups:
#     setosa versicolor  virginica 
#  0.3333333  0.3333333  0.3333333 

# Group means:
#            Petal.Length
# setosa            1.462
# versicolor        4.260
# virginica         5.552

# Coefficients of linear discriminants:
#                   LD1
# Petal.Length 2.323774

Upvotes: 3

Metrics
Metrics

Reputation: 15458

myForm<-as.formula(paste("Species","Petal.Length",sep="~"))
lgrIris<-glm(formula=myForm, data=iris, family="binomial")

Upvotes: 1

Related Questions