K-Means Clustering Algorithm
সম্পূর্ণ বাংলা লেকচার নোট

# প্রয়োজনীয় লাইব্রেরি
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.cluster import KMeans
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import silhouette_score
import warnings
warnings.filterwarnings('ignore')
plt.style.use('seaborn-v0_8')
np.random.seed(42)

# Mall Customer Dataset তৈরি
n = 200
df = pd.DataFrame({
    'CustomerID':    range(1, n+1),
    'Age':           np.random.randint(18, 70, n),
    'Annual_Income': np.random.randint(15, 140, n),  # হাজার টাকায়
    'Spending_Score':np.random.randint(1,  100, n)
})
print("Dataset আকার:", df.shape)
print(df.head())

print("Dataset আকার:", df.shape)
print(df.head())

print(df.isnull().sum())
print("\nDuplicate rows:", df.duplicated().sum())

print(df.describe().round(2))

# Scatter Plot: আয় vs Spending Score
plt.figure(figsize=(10, 5))
plt.subplot(1,2,1)
plt.scatter(df['Annual_Income'], df['Spending_Score'],
            color='#534AB7', alpha=0.7, s=50)
plt.xlabel('বার্ষিক আয় (হাজার)')
plt.ylabel('Spending Score')
plt.title('Clustering আগে — Unlabeled')

plt.subplot(1,2,2)
plt.scatter(df['Age'], df['Spending_Score'],
            color='#993C1D', alpha=0.7, s=50)
plt.xlabel('বয়স')
plt.ylabel('Spending Score')
plt.title('বয়স vs Spending')
plt.tight_layout()
plt.show()

# Clustering-এর জন্য Feature বেছে নেওয়া
X = df[['Annual_Income', 'Spending_Score']]

# Scaling করার আগে ও পরে তুলনা
print("Scaling-এর আগে:")
print(X.describe().round(2))

scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)

print("\nScaling-এর পরে (প্রথম ৫ সারি):")
print(pd.DataFrame(X_scaled, columns=X.columns).head().round(3))

# WCSS হিসাব করি K=1 থেকে K=10
wcss = []
for k in range(1, 11):
    km = KMeans(n_clusters=k, random_state=42, n_init=10)
    km.fit(X_scaled)
    wcss.append(km.inertia_)
    print(f"K={k}: WCSS={km.inertia_:.2f}")

# Elbow Curve
plt.figure(figsize=(9, 5))
plt.plot(range(1, 11), wcss,
         marker='o', linewidth=2.5,
         color='#534AB7', markerfacecolor='#E24B4A', markersize=9)
plt.axvline(x=3, color='#E24B4A', linestyle='--', linewidth=2,
            label='Elbow Point (K=3)')
plt.xlabel('K (Cluster সংখ্যা)', fontsize=12)
plt.ylabel('WCSS', fontsize=12)
plt.title('Elbow Method — সেরা K নির্বাচন', fontsize=14)
plt.xticks(range(1, 11))
plt.legend()
plt.grid(True, alpha=0.4)
plt.tight_layout()
plt.show()
print("\n✅ Elbow দেখে বোঝা যাচ্ছে K=3 সেরা!")

# K=3 দিয়ে K-Means মডেল Train
km3 = KMeans(n_clusters=3, random_state=42, n_init=10)
df['Cluster'] = km3.fit_predict(X_scaled)

print("✅ Clustering সম্পন্ন!")
print("\nCluster Distribution:")
print(df['Cluster'].value_counts().sort_index())

print("\nCentroid অবস্থান (Scaled):")
print(pd.DataFrame(km3.cluster_centers_,
      columns=['Annual_Income','Spending_Score']).round(3))

# Cluster চিত্র
colors = ['#534AB7','#E24B4A','#3B6D11']
labels = ['Low Income High Spender', 'High Income Low Spender',
           'Average Customer']

plt.figure(figsize=(10, 6))
for i in range(3):
    mask = df['Cluster'] == i
    plt.scatter(df[mask]['Annual_Income'],
                df[mask]['Spending_Score'],
                c=colors[i], label=labels[i], s=60, alpha=0.8)

# Centroid গুলো দেখানো
centroids_orig = scaler.inverse_transform(km3.cluster_centers_)
plt.scatter(centroids_orig[:,0], centroids_orig[:,1],
            marker='*', s=300, c='black', zorder=5, label='Centroids')

plt.xlabel('বার্ষিক আয় (হাজার টাকা)', fontsize=12)
plt.ylabel('Spending Score', fontsize=12)
plt.title('Mall Customer Segmentation — K=3', fontsize=14)
plt.legend(bbox_to_anchor=(1.05,1), loc='upper left', fontsize=10)
plt.tight_layout()
plt.show()

# Silhouette Score হিসাব
sil_score = silhouette_score(X_scaled, df['Cluster'])
print(f"✅ Silhouette Score: {sil_score:.4f}")
print("(১-এর কাছাকাছি মানে ভালো Clustering)")

# বিভিন্ন K-তে Silhouette Score তুলনা
print("\nK মান | Silhouette Score")
print("-" * 28)
for k in range(2, 9):
    km_t = KMeans(n_clusters=k, random_state=42, n_init=10)
    labels_t = km_t.fit_predict(X_scaled)
    sc = silhouette_score(X_scaled, labels_t)
    print(f"K={k}   | {sc:.4f}")