ass11(k-means)

import pandas as pd 

import numpy as np 

import matplotlib.pyplot as plt 

from sklearn.preprocessing import StandardScaler 

from sklearn.cluster import KMeans 

from sklearn.decomposition import PCA 

from mpl_toolkits.mplot3d import Axes3D 

from scipy.spatial.distance import cdist 

# Load the dataset (update name if different) 

df = pd.read_csv("customer_segmentation.csv") 

df.head() 

df = df.dropna() 

X = df.select_dtypes(include=[np.number]) 

# Standardize the features 

scaler = StandardScaler() 

X_scaled = scaler.fit_transform(X) 

kmeans_euc = KMeans(n_clusters=3, random_state=42) 

labels_euc = kmeans_euc.fit_predict(X_scaled) 

df['Cluster_Euclidean'] = labels_euc 

def kmeans_manhattan(X, k, max_iters=100): 

np.random.seed(42) 

centroids = X[np.random.choice(range(len(X)), k, replace=False)] 

for _ in range(max_iters): 

        labels = np.argmin(cdist(X, centroids, metric='cityblock'), axis=1) 

        new_centroids = np.array([X[labels == i].mean(axis=0) for i in range(k)]) 

        if np.allclose(centroids, new_centroids): 

            break 

        centroids = new_centroids 

 

    return labels, centroids 

 

labels_man, _ = kmeans_manhattan(X_scaled, k=3) 

df['Cluster_Manhattan'] = labels_man 

 

pca = PCA(n_components=2) 

X_pca = pca.fit_transform(X_scaled) 

 

plt.figure(figsize=(12,5)) 

 

# Euclidean Clustering 

plt.subplot(1,2,1) 

plt.scatter(X_pca[:,0], X_pca[:,1], c=df['Cluster_Euclidean'], cmap='viridis') 

plt.title('KMeans (Euclidean Distance)') 

plt.xlabel('PCA 1') 

plt.ylabel('PCA 2') 

# Manhattan Clustering 

plt.subplot(1,2,2) 

plt.scatter(X_pca[:,0], X_pca[:,1], c=df['Cluster_Manhattan'], cmap='plasma') 

plt.title('KMeans (Manhattan Distance)') 

plt.xlabel('PCA 1') 

plt.ylabel('PCA 2') 

plt.tight_layout() 

plt.show() 

pca_3d = PCA(n_components=3) 

X_3d = pca_3d.fit_transform(X_scaled) 

fig = plt.figure(figsize=(10, 6)) 

ax = fig.add_subplot(111, projection='3d') 

ax.scatter(X_3d[:, 0], X_3d[:, 1], X_3d[:, 2], c=df['Cluster_Euclidean'], 

cmap='coolwarm') 

ax.set_title('3D Clustering (Euclidean)') 

ax.set_xlabel("PC1") 

ax.set_ylabel("PC2") 

ax.set_zlabel("PC3") 

plt.show()