One of the best post for studying machine learning with R

This is definitely one of the best link to learn about machine learning using R.

Although you can follow through the blog, there are some R syntax that you might not be so familiar with :-

a) What is the purpose of this command line

names(iris) <- c="" code="" epal.length="" epal.width="" etal.length="" etal.width="" pecies="">

This is just for naming your dataset. For example, lets create a list of bird and their size

bird <- bird1="1," bird2="5)</p" list="">
Size 1, would probably be sparrow and size 5 would be a larger bird call eagle. Let change this, using names

names(bird)[1] <- p="" sparrow="">names(bird)[2] <- eagle="" nbsp="" p="">
Output

> bird
$sparrow
[1] 1

$eagle
[1] 5


b) Funny looking operator "%>%"

This is a pipe operator import from magrittr. which is basically allow us to invoke command in sequence.

iris %>% ggvis(~Sepal.Length, ~Sepal.Width, fill = ~Species) %>% layer_points()

This is the same as calling

ggvis(~Sepal.Length, ~Sepal.Width, fill = ~Species) 

then

layer_points()

c) Another important aspect in the tutorial is that we divide our dataset into training and test using a sample command.

ind <- 0.33="" code="" iris="" nrow="" prob="c(0.67," replace="TRUE," sample="">

iris.training <- 1:4="" code="" ind="=2," iris.test="" iris="">

Notice that knn function requires this input parameters.

knn(train, test, cl, k = 1, l = 0, prob = FALSE, use.all = TRUE)

Some description of the parameters required.

train = input for our training set

test = input for our training set

cl = factor for true classification for training set

k = neighbor considered  We have 3 distinct neighbour for our dataset.




Other side notes.

How to install R package

install.packages("")

Check if you have a package installed

any(grepl("", installed.packages()))