Scribd
Upload
2 views

Assignment4

The document is a Jupyter notebook that implements the K-means clustering algorithm using Python libraries such as pandas and numpy. It reads data from a text file, initializes random centroids, and iteratively assigns data points to clusters based on distance until convergence. Finally, it visualizes the clustered data points using matplotlib.

Uploaded by

sadiamohiuddin3009
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

Assignment4

The document is a Jupyter notebook that implements the K-means clustering algorithm using Python libraries such as pandas and numpy. It reads data from a text file, initializes random centroids, and iteratively assigns data points to clusters based on distance until convergence. Finally, it visualizes the clustered data points using matplotlib.

Uploaded by

sadiamohiuddin3009
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

8/20/22, 7:33 PM 180104072_Assignment4.

ipynb - Colaboratory

import pandas as pd
import numpy as np
import matplotlib
import matplotlib.pyplot as plt

from google.colab import drive

drive.mount('/content/gdrive')

Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive

main_directory = '/content/gdrive/MyDrive/Pattern Lab/Assignment4/data_k_mean.txt'


train = pd.read_csv(main_directory, sep=' ' , header = None)
train1 = train.to_numpy()
print(train1)

[[-7.87157 -4.86573]
[-4.76661 -6.87944]
[-6.67986 -5.8308 ]
...
[ 6.91832 -0.32132]
[-8.23828 -4.00405]
[-5.75112 -5.99531]]

#plotting all data points


plt.scatter(train[0], train[1], c = 'blue', marker = 's')
plt.show()

#taking k as input
k = int(input("Enter the value of k : "))

Enter the value of k : 2

#random centroids for 1st iteration


np.random.seed(seed=72)
random_numbers = np.random.randint(low=0, high=len(train1), size=(k))
https://colab.research.google.com/drive/1A2rXv6R5N4bVbrStI9bsxdhKsJlzNVYu?authuser=3#scrollTo=QzW8dbY5YkR0&printMode=true 1/3
8/20/22, 7:33 PM 180104072_Assignment4.ipynb - Colaboratory

centroids = [train1[random_numbers[i]] for i in range(k)]


print(centroids)

[array([ 6.80375, -0.13017]), array([6.68468, 0.85224])]

distance = [] #to store the distance from point to classes


index_clusters = [-1 for i in range(len(train1))] #to store class corresponding to index
count = 0 #to count the iteration number
clusters = {} #to store class numbers as keys and data points as values
for x in range(500):
count = x
# flag to keep track whether change occurs or not
flag = 0
for y in range(k):
clusters[y] = []
# iterate through each data points
for i in range(len(train1)):
distance = []
for j in range(k):
dist = np.sqrt(pow(abs((train1[i][0] - centroids[j][0])), 2) + pow(abs((train1
distance.append(dist)
index = distance.index(min(distance))
# check whether the change occurs or not
if index_clusters[i] != index:
flag = 1
index_clusters[i] = index
clusters[index].append(train1[i])
# if change occurs
if flag == 0:
break
# calculating new centroids
centroids = [np.mean(np.asarray(clusters[z]), axis=0) for z in range(k)]

x1 = np.asarray(clusters[0])[:, 0]
y1 = np.asarray(clusters[0])[:, 1]

x2 = np.asarray(clusters[1])[:, 0]
y2 = np.asarray(clusters[1])[:, 1]

# plotting classified data points of two classes with different colored marker
plt.scatter(x1, y1, c = 'red', marker = 's', label = 'Class 1')
plt.scatter(x2, y2, c = 'green', marker = 'H', label = 'Class 2')
plt.legend(loc = 'best')
plt.show()

https://colab.research.google.com/drive/1A2rXv6R5N4bVbrStI9bsxdhKsJlzNVYu?authuser=3#scrollTo=QzW8dbY5YkR0&printMode=true 2/3
8/20/22, 7:33 PM 180104072_Assignment4.ipynb - Colaboratory

check 0s completed at 7:32 PM

https://colab.research.google.com/drive/1A2rXv6R5N4bVbrStI9bsxdhKsJlzNVYu?authuser=3#scrollTo=QzW8dbY5YkR0&printMode=true 3/3

You might also like