clustering_ver2.R

# CLUSTERING IRIS.R
#
#
# Version:
#
# Date:
# Author:
#
#TOC> ==========================================================================
#TOC>
#TOC>   Section  Title                                            Line
#TOC> --------------------------------------------------------------------------
#TOC>   1        Packages
#TOC>   2        Dataset exploration
#TOC>   3        Clustering
#TOC>   4        Multidimensional clustering
#TOC> ==========================================================================

# ====    1  Packages  =========================================================

if (!require(ggplot2, quietly = TRUE)) {
    install.packages("ggplot2")
    library(ggplot2)
}

# ====    2 Dataset exploration  ===============================================

my_data <- iris

# Let's explore the dataset
str(my_data)

summary(my_data$Petal.Length)

# Let's explore the distribution of petal length across the different species

qplot(
    Petal.Length,
    data = my_data,
    geom = 'density',
    color = Species,
    fill = Species
)

# ====    3 Clustering  ========================================================

# Can we exctract the different species into clusters using petal lenght?

my_data2 <- my_data[, "Petal.Length"]

num_clusters <- 3
set.seed(1234)
result <- kmeans(my_data2, num_clusters, nstart = 20)

# Let's see how the original species have been assigned into clusters

table(result$cluster, my_data$Species)

# How many measurements have been misclassified?

# Let's plot sepal lenght per cluster

qplot(Petal.Length,
      data = my_data,
      color = factor(result$cluster))

# ====    4 Multidimensional clustering  =======================================

# Can we exctract the different species into clusters using 2 attributes?

my_data3 <- my_data[, c("Petal.Width", "Petal.Length")]

num_clusters <- 3
set.seed(1234)
result <- kmeans(my_data3, num_clusters, nstart = 20)

table(result$cluster, my_data$Species)

# Let's plot sepal lenght and petal width per cluster

qplot(Petal.Length,
      Petal.Width,
      data = my_data,
      color = factor(result$cluster))