Name of Student: Ahmed Mobin Ahmed Shaikh

Roll Number: 45 Lab Practical Number: 07

Title of Lab Assignment: Implementation and analysis of clustering

algorithms like K-Means, K-medoid.

DOP: 06/03/24 DOS: 06/03/24

CO Mapped: PO Mapped: Signature:

CO2 PO3, PO5,
PO6, PO7,
PO11,
PO12
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory

keyboard_arrow_down 1. K-Means Clustering:

I. IMPORTING DATASET:

import numpy as nm
import matplotlib.pyplot as mtp
import pandas as pd

# Importing the dataset

dataset = pd.read_csv('F1Drivers_Dataset.csv')
#dataset
dataset

Driver Nationality Seasons Championships Race_Entries Race_Starts Pole_P

Carlo [1962,
0 Italy 0.0 3.0 0.0
Abate 1963]

George United [1951,

1 0.0 2.0 2.0
Abecassis Kingdom 1952]

Kenny United [1983,

2 0.0 10.0 3.0
Acheson Kingdom 1985]

[1968,
Andrea 1970,
3 de Italy 1971, 0.0 36.0 30.0
Adamich 1972,
1973]

Philippe
4 Belgium [1994] 0.0 2.0 2.0
Adams

... ... ... ... ... ... ...

Emilio
863 Spain [1976] 0.0 1.0 0.0
Zapico

Zhou
864 China [2022] 0.0 23.0 23.0
Guanyu

[1999,
2000,
Ricardo
865 Brazil 2001, 0.0 37.0 36.0
Zonta
2004,
2005]

[1975,
Renzo
866 Italy 1976, 0.0 7.0 7.0
Zorzi
1977]

[1979,
Ricardo
867 Argentina 1980, 0.0 11.0 10.0
Zunino
1981]

868 rows × 22 columns

keyboard_arrow_down EXTRACTING INDEPENDENT VARIABLES:

SOURCE CODE:

x = dataset[['Race_Entries', 'Race_Starts']]
#finding optimal number of clusters using the elbow method
from sklearn.cluster import KMeans
wcss_list= [] #Initializing the list for the values of WCSS
#Using for loop for iterations from 1 to 10.
for i in range(1, 11):
kmeans = KMeans(n_clusters=i, init='k-means++', max_iter=300, n_init=10, random_state=0)
kmeans.fit(x)
wcss_list.append(kmeans.inertia_)
mtp.plot(range(1, 11), wcss_list)
mtp.title('The Elbow Method Graph')
mtp.xlabel('Number of clusters(k)')
mtp.ylabel('wcss_list')
mtp.show()

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 1/7
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory

keyboard_arrow_down TRAINING K-MEANS MODEL ON A DATASET:

SOURCE CODE:

# Applying KMeans with 5 clusters

kmeans = KMeans(n_clusters=5, init='k-means++', random_state=42)
y_predict = kmeans.fit_predict(x)

# Visualizing the clusters

mtp.scatter(x[y_predict == 0]['Race_Entries'], x[y_predict == 0]['Race_Starts'], s=100, c='blue', label='Cluster 1')
mtp.scatter(x[y_predict == 1]['Race_Entries'], x[y_predict == 1]['Race_Starts'], s=100, c='green', label='Cluster 2')
mtp.scatter(x[y_predict == 2]['Race_Entries'], x[y_predict == 2]['Race_Starts'], s=100, c='red', label='Cluster 3')
mtp.scatter(x[y_predict == 3]['Race_Entries'], x[y_predict == 3]['Race_Starts'], s=100, c='cyan', label='Cluster 4')
mtp.scatter(x[y_predict == 4]['Race_Entries'], x[y_predict == 4]['Race_Starts'], s=100, c='magenta', label='Cluster 5')

# Plotting centroids
mtp.scatter(kmeans.cluster_centers_[:, 0], kmeans.cluster_centers_[:, 1], s=300, c='yellow', label='Centroid')

mtp.title('Clusters of Driver data')

mtp.xlabel('Race_Entries')
mtp.ylabel('Race_Starts')
mtp.legend()
mtp.show()

/usr/local/lib/python3.10/dist-packages/sklearn/cluster/_kmeans.py:870: FutureWarning
warnings.warn(

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 2/7
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory

keyboard_arrow_down 2. K-Medoids:
I. Importing Packages & Loading Dataset:

SOURCE CODE:

!pip install scikit-learn-extra

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from sklearn_extra.cluster import KMedoids
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import silhouette_score
from sklearn import datasets, metrics

# Load the Wine dataset

wine = datasets.load_wine()
x = wine.data # Features
y = wine.target # Target labels
wine

Collecting scikit-learn-extra
Downloading scikit_learn_extra-0.3.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.0 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.0/2.0 MB 8.9 MB/s eta 0:00:00
Requirement already satisfied: numpy>=1.13.3 in /usr/local/lib/python3.10/dist-packages (from scikit-learn-extra) (1.25.2)
Requirement already satisfied: scipy>=0.19.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn-extra) (1.11.4)
Requirement already satisfied: scikit-learn>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn-extra) (1.2.2)
Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn>=0.23.0->scikit-learn-
Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn>=0.23.0->scikit
Installing collected packages: scikit-learn-extra
Successfully installed scikit-learn-extra-0.3.0
{'data': array([[1.423e+01, 1.710e+00, 2.430e+00, ..., 1.040e+00, 3.920e+00,
1.065e+03],
[1.320e+01, 1.780e+00, 2.140e+00, ..., 1.050e+00, 3.400e+00,
1.050e+03],
[1.316e+01, 2.360e+00, 2.670e+00, ..., 1.030e+00, 3.170e+00,
1.185e+03],
...,
[1.327e+01, 4.280e+00, 2.260e+00, ..., 5.900e-01, 1.560e+00,
8.350e+02],
[1.317e+01, 2.590e+00, 2.370e+00, ..., 6.000e-01, 1.620e+00,
8.400e+02],
[1.413e+01, 4.100e+00, 2.740e+00, ..., 6.100e-01, 1.600e+00,
5.600e+02]]),
'target': array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2,
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
2, 2]),
'frame': None,
'target_names': array(['class_0', 'class_1', 'class_2'], dtype='<U7'),
'DESCR': '.. _wine_dataset:\n\nWine recognition dataset\n------------------------\n\n**Data Set Characteristics:**\n\n
:Number of Instances: 178\n :Number of Attributes: 13 numeric, predictive attributes and the class\n :Attribute
Information:\n \t\t- Alcohol\n \t\t- Malic acid\n \t\t- Ash\n\t\t- Alcalinity of ash \n \t\t- Magnesium\n\t\t- Total phenols\n
\t\t- Flavanoids\n \t\t- Nonflavanoid phenols\n \t\t- Proanthocyanins\n\t\t- Color intensity\n \t\t- Hue\n \t\t- OD280/OD315 of
diluted wines\n \t\t- Proline\n\n - class:\n - class_0\n - class_1\n - class_2\n\t\t\n
:Summary Statistics:\n \n ============================= ==== ===== ======= =====\n Min
Max Mean SD\n ============================= ==== ===== ======= =====\n Alcohol: 11.0 14.8
13.0 0.8\n Malic Acid: 0.74 5.80 2.34 1.12\n Ash: 1.36 3.23 2.36
0.27\n Alcalinity of Ash: 10.6 30.0 19.5 3.3\n Magnesium: 70.0 162.0 99.7 14.3\n
Total Phenols: 0.98 3.88 2.29 0.63\n Flavanoids: 0.34 5.08 2.03 1.00\n
Nonflavanoid Phenols: 0.13 0.66 0.36 0.12\n Proanthocyanins: 0.41 3.58 1.59 0.57\n Colour
Intensity: 1.3 13.0 5.1 2.3\n Hue: 0.48 1.71 0.96 0.23\n OD280/OD315 of
diluted wines: 1.27 4.00 2.61 0.71\n Proline: 278 1680 746 315\n
============================= ==== ===== ======= =====\n\n :Missing Attribute Values: None\n :Class Distribution: class_0
(59), class_1 (71), class_2 (48)\n :Creator: R.A. Fisher\n :Donor: Michael Marshall (MARSHALL%[email protected])\n
:Date: July, 1988\n\nThis is a copy of UCI ML Wine recognition datasets.\nhttps://archive.ics.uci.edu/ml/machine-learning-
databases/wine/wine.data\n\nThe data is the results of a chemical analysis of wines grown in the same\nregion in Italy by three
different cultivators. There are thirteen different\nmeasurements taken for different constituents found in the three types
of\nwine.\n\nOriginal Owners: \n\nForina, M. et al, PARVUS - \nAn Extendible Package for Data Exploration, Classification and
Correlation. \nInstitute of Pharmaceutical and Food Analysis and Technologies,\nVia Brigata Salerno, 16147 Genoa,
Italy.\n\nCitation:\n\nLichman, M. (2013). UCI Machine Learning Repository\n[https://archive.ics.uci.edu/ml]. Irvine, CA:
University of California,\nSchool of Information and Computer Science. \n\n.. topic:: References\n\n (1) S. Aeberhard, D.
Coomans and O. de Vel, \n Comparison of Classifiers in High Dimensional Settings, \n Tech. Rep. no. 92-02, (1992), Dept. of

keyboard_arrow_down II. Scaling and Fitting KMedoids:

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 3/7
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory
SOURCE CODE:

# Scaling the features

scaler = StandardScaler().fit(x)
x_scaled = scaler.transform(x)

# II. Scaling and Fitting KMedoids

kMedoids = KMedoids(n_clusters=3, random_state=0)
kMedoids.fit(x_scaled)
y_kmed = kMedoids.predict(x_scaled)

keyboard_arrow_down III. Silhouette Method to evaluate cluster:

SOURCE CODE:

silhouette_avg = silhouette_score(x_scaled, y_kmed)

print("Silhouette Score:", silhouette_avg)

Silhouette Score: 0.26597740204536796

keyboard_arrow_down IV. Silhouette Width to find number of cluster:

SOURCE CODE:

sw = []
for i in range(2, 11):
kMedoids = KMedoids(n_clusters=i, random_state=0)
kMedoids.fit(x_scaled)
y_kmed = kMedoids.predict(x_scaled)
silhouette_avg = silhouette_score(x_scaled, y_kmed)
sw.append(silhouette_avg)

plt.plot(range(2, 11), sw)

plt.title('Silhouette Score')
plt.xlabel('Number of clusters')
plt.ylabel('Silhouette Width')
plt.show()

keyboard_arrow_down V. Computing Purity:

SOURCE CODE:

def purity_score(y_true, y_pred):

contingency_matrix = metrics.cluster.contingency_matrix(y_true, y_pred)
return np.sum(np.amax(contingency_matrix, axis=0)) / np.sum(contingency_matrix)

print("Purity Score:", purity_score(y, y_kmed))

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 4/7
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory

Purity Score: 0.898876404494382

keyboard_arrow_down VI. How extreme values effect K-Medoid compared to K-means:

SOURCE CODE:

kmeans = KMeans(n_clusters=3, init='random', max_iter=300, n_init=10, random_state=0)

y_kmeans = kmeans.fit_predict(x_scaled)
print("Purity Score for K-Means:", purity_score(y, y_kmeans))

Purity Score for K-Means: 0.9662921348314607

keyboard_arrow_down VII. Plotting values:

SOURCE CODE:

plt.scatter(x_scaled[y_kmeans == 0, 0], x_scaled[y_kmeans == 0, 1], s=100, c='red', label='Cluster 1')

plt.scatter(x_scaled[y_kmeans == 1, 0], x_scaled[y_kmeans == 1, 1], s=100, c='blue', label='Cluster 2')
plt.scatter(x_scaled[y_kmeans == 2, 0], x_scaled[y_kmeans == 2, 1], s=100, c='green', label='Cluster 3')
plt.scatter(kmeans.cluster_centers_[:, 0], kmeans.cluster_centers_[:, 1], s=300, c='yellow', label='Centroids')
plt.legend()
plt.show()

keyboard_arrow_down VIII. Adding extreme values:

SOURCE CODE:

import numpy as np

# Add extreme values to your dataset

extreme_values = np.array([[10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10],
[15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15],
[12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12]])
m = np.append(x, extreme_values, axis=0)
y_extreme = np.append(y, [2, 2, 2])
print(y_extreme)
print("we see 3 observations are added over here.-", m.shape)

[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2]
we see 3 observations are added over here.- (181, 13)

SOURCE CODE:

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 5/7
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory

scaler = StandardScaler().fit(m)
x_scaled_extreme = scaler.transform(m)

# Perform KMeans clustering

kmeans_extreme = KMeans(n_clusters=3, init='random', max_iter=300, n_init=10, random_state=0)
y_kmeans_extreme = kmeans_extreme.fit_predict(x_scaled_extreme)

# Calculate purity score

purity_extreme = purity_score(y_extreme, y_kmeans_extreme)
print(purity_extreme)

0.7016574585635359

keyboard_arrow_down IX. Plot:

SOURCE CODE:

plt.scatter(x_scaled_extreme[y_kmeans_extreme == 0, 0], x_scaled_extreme[y_kmeans_extreme == 0, 1], s=100, c='red', label='C1')

plt.scatter(x_scaled_extreme[y_kmeans_extreme == 1, 0], x_scaled_extreme[y_kmeans_extreme == 1, 1], s=100, c='blue', label='C2')
plt.scatter(x_scaled_extreme[y_kmeans_extreme == 2, 0], x_scaled_extreme[y_kmeans_extreme == 2, 1], s=100, c='green', label='C3')
plt.scatter(kmeans_extreme.cluster_centers_[:, 0], kmeans_extreme.cluster_centers_[:, 1], s=100, c='yellow', label='Centroids')
plt.legend()
plt.show()

output

SOURCE CODE:

data = [['k-Means', 0.81], ['k-Means with Outliers', purity_extreme],

['k-Medoid', 0.84], ['K Medoid with outliers', 0.86]]
df = pd.DataFrame(data, columns=['Method', 'Purity'])
df.plot.bar(x='Method', y='Purity', title='Cluster Quality')
plt.show()

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 6/7
3/8/24, 9:38 PM 45B_AIML_Practical07_Clustering.ipynb - Colaboratory

https://colab.research.google.com/drive/1lMbyWJ57Y8yEjDX02i2XrF_nKrKep-Fk#scrollTo=Z7Y3hDiCyg3e&printMode=true 7/7