DBSCAN: Core Concepts and Definitions
What you'll learn: You'll understand the four key building blocks that make DBSCAN work: epsilon neighborhoods, core points, border points, and noise points.
The Foundation of Density-Based Clustering
DBSCAN (Density-Based Spatial Clustering of Applications with Noise) identifies clusters by finding regions where data points are tightly packed together. Unlike K-Means, which uses distance from centroids, DBSCAN looks at how densely points are concentrated in neighborhoods.
The Four Essential Concepts
Epsilon (ε) Neighborhood: Imagine drawing a circle of radius ε around each point. Every point within that circle is considered a "neighbor." The epsilon parameter defines how large this neighborhood is.
Core Points: A point is a "core point" if it has at least min_samples neighbors within its ε-neighborhood (including itself). Think of core points as the heart of dense regions—they have enough friends nearby to anchor a cluster.
Border Points: These points fall within the ε-neighborhood of a core point but don't have enough neighbors themselves to be core points. They're on the edge of clusters—part of the crowd, but not at its center.
Noise Points: Points that are neither core points nor border points are classified as noise (outliers). They're isolated stragglers that don't belong to any dense region.
A Simple Analogy
Think of a busy city: core points are downtown areas with many people clustered together. Border points are the suburbs—close enough to downtown to be part of the city, but less densely populated. Noise points are isolated farmhouses far from any urban area.
Key Takeaway: DBSCAN defines clusters using four concepts: epsilon neighborhoods determine who's nearby, core points anchor dense regions, border points sit at cluster edges, and noise points are outliers rejected by all clusters.