What is the Silhouette Score?

The Silhouette Score is a metric used to evaluate the quality of clustering. It measures how similar a data point is to its own cluster compared to other clusters.

Why is Silhouette Score Important?

While the Elbow Method helps choose K, the Silhouette Score helps evaluate the actual quality of the clusters.

• Measures how well data points are grouped.
• Detects overlapping or poorly defined clusters.
• Helps validate K-Means, Hierarchical, and other clustering algorithms.
• Provides a single numeric value between -1 and +1.
• Works even when ground truth labels are missing.

Concept Behind Silhouette Score

The Silhouette Score uses two key distances:

Mathematical Formula

Where:

• S(i) = Silhouette score of point i
• a(i) = mean distance to points in the same cluster
• b(i) = mean distance to points in the nearest different cluster

Silhouette Score Range

How Silhouette Score Works

Visual Intuition

Python Example — Silhouette Score with K-Means

pip install numpy matplotlib scikit-learn

import numpy as np
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
from sklearn.metrics import silhouette_score
from sklearn.datasets import make_blobs

# Generate sample data
X, _ = make_blobs(n_samples=300, centers=4, cluster_std=0.7, random_state=42)

# Compute Silhouette Scores for different K
silhouette_scores = []
for k in range(2, 11):
    km = KMeans(n_clusters=k, init="k-means++", random_state=42, n_init=10)
    labels = km.fit_predict(X)
    score = silhouette_score(X, labels)
    silhouette_scores.append(score)
    print(f"K={k}, Silhouette Score={score:.3f}")

# Plot results
plt.plot(range(2, 11), silhouette_scores, marker="o", color="green")
plt.title("Silhouette Score vs K")
plt.xlabel("Number of Clusters (K)")
plt.ylabel("Silhouette Score")
plt.grid(True)
plt.show()

Silhouette Plot

Besides the score, you can use a silhouette plot to analyze each cluster's quality.

from sklearn.metrics import silhouette_samples
import matplotlib.cm as cm

km = KMeans(n_clusters=4, random_state=42, n_init=10)
labels = km.fit_predict(X)

sample_silhouette = silhouette_samples(X, labels)

y_lower = 10
for i in range(4):
    cluster_silhouette = sample_silhouette[labels == i]
    cluster_silhouette.sort()
    size = cluster_silhouette.shape[0]
    y_upper = y_lower + size
    color = cm.viridis(float(i) / 4)
    plt.fill_betweenx(np.arange(y_lower, y_upper),
                      0, cluster_silhouette, facecolor=color, alpha=0.7)
    y_lower = y_upper + 10

plt.title("Silhouette Plot")
plt.xlabel("Silhouette Coefficient")
plt.ylabel("Cluster Index")
plt.axvline(x=np.mean(sample_silhouette), color="red", linestyle="--")
plt.show()

Elbow Method vs Silhouette Score

Real-Life Analogy

Real-World Applications

Advantages

• Provides a single numerical value.
• Works on unlabeled data.
• Measures both cohesion and separation.
• Easy to interpret (-1 to +1).
• Helps validate clustering algorithms.

Disadvantages

Common Mistakes to Avoid

Best Practices

Silhouette vs Other Clustering Metrics

Importance of Silhouette Score

Golden Rule