🦠 DBSCAN

DBSCAN is a clustering algorithm based on Density.

👥 The Crowd Analogy

Unlike K-Means (which draws rigid circles), DBSCAN looks for organic "crowds" of data points. It doesn't require you to guess K!

🐍 Python Implementation

from sklearn.cluster import DBSCAN

X = [[1, 2], [1, 4], [10, 2], [10, 4], [100, 100]] # the last one is an outlier!

# eps = maximum distance between two points to be considered "friends"
dbscan = DBSCAN(eps=3, min_samples=2)
labels = dbscan.fit_predict(X)

print("Cluster labels:", labels)
# Notice the outlier gets a label of "-1" (Noise!)