Skip to content

How to do Naive Bayes in R?

4 messages · Tim LIU, Dylan Beaudette, Spencer Graves

Tim LIU
# 
I am wondering if anybody here have a simple example in R for Naive 
Bayes.

For example, I can do k-means clustering on the "iris" data -

data(iris)
cl <- kmeans(iris[,1:4], 3)
cl$cluster
cbind(1:150,iris$Species)

===========

But how to do Naive Bayes classification in the same "iris" data?

Many thanks!
Dylan Beaudette
#
On Wednesday 06 May 2009, Tim LIU wrote:
# start here:
help.search('bayes')

# in which you will find:
e1071::naiveBayes

# which means:
library(e1071)
?naiveBayes

# and compared to your example:
cl <- kmeans(iris[,1:4], 3)
table(cl$cluster, iris[,5])

m <- naiveBayes(iris[,1:4], iris[,5])
table(predict(m, iris[,1:4]), iris[,5])

Dylan
Spencer Graves
# 
help.search('bayes') only searches installed packages. 

      To go beyond that, you might try the following: 


library(RSiteSearch)

bayes <- RSiteSearch.function('bayes', 999)
summary(bayes)
# This produces a Parato analysis of packages in terms of references
# to the search term in their help pages. 
# In this example, it does NOT find 'e1071'. 

n.b <- RSiteSearch.function('naive Bayes')
HTML(n.b)
# This displays a table in a web browser sorted by package and "Score"
# with links to the help pages via the web. 


      Beyond this, "www.r-project.org -> Documentation:  Books" provides 
a list of books about various aspects of R, including several references 
to "Bayes". 


      Hope this helps. 
      Spencer
Dylan Beaudette wrote:
Dylan Beaudette
#
On Wednesday 06 May 2009, spencerg wrote:
Thanks for the clarification. 

Dylan