Clustering (K-Means)

Let's Think About This for a Second

Clustering is an Unsupervised Learning technique that automatically groups unlabeled data into groups (clusters) of similar characteristics — you don't need to give the model the 'correct answer' (label) from outside; the data finds its own patterns. K-Means is one of the most commonly used clustering algorithms — you have to specify 'K' (the number of clusters) in advance, and the algorithm assigns each data point to its 'nearest cluster center,' then iteratively adjusts the cluster centers based on the data average (until convergence).

Connecting to a Real Scenario

Clustering e-commerce customer data (purchase amount, frequency) with K-Means without any labels lets you automatically discover customer segments straight from the data, like 'high-value frequent buyer,' 'occasional buyer,' and 'one-time buyer' (the marketing team can use these segment results for personalized campaigns) — writing `from sklearn.cluster import KMeans; model = KMeans(n_clusters=3); model.fit(X)` assigns each customer into cluster 0/1/2.

• Clustering = grouping unlabeled data into similar groups (clusters) — Unsupervised Learning
• K-Means = you set the number of clusters (K) in advance, then iteratively adjust the cluster centers
• Clustering is commonly used for customer segmentation and anomaly detection

Let's Look at It Together

from sklearn.cluster import KMeans

# X here has no labels — just features like [purchase_amount, frequency]
model = KMeans(n_clusters=3, random_state=42)
model.fit(X)

clusters = model.labels_
print(clusters[:10])  # which cluster (0, 1, or 2) each customer belongs to
print(model.cluster_centers_)  # the "average" customer per cluster

[0 2 1 0 0 2 1 1 0 2]
[[ 150.2   3.1]
 [ 800.5  12.4]
 [  45.8   1.2]]