DBSCAN (Density-Based Spatial Clustering of Applications with Noise

In this lecture, I learned about the Unsupervised Machine Learning Algorithm DBSCAN (Density-Based Spatial Clustering Application with Noise). Here, "Density" means the Number of data points in some area. "Spatial" means density in some space. "Noise" means the data points that are not part of any cluster. DBSCAN has 3 types of points: 1. Core points 2. Boarder points 3. Outlier points / Noise. In DBSCAN, we have to give 2 values: 1. Epsilone "E" (radius of a circle) 2. MinPts (Min number of points in a circle)

DBSCAN: Density(something in some space: Data points / Area) base spatial(space) clustering(divide data [oint in group) application noise(data points outside selected area).eps(epsolon: radius of circle, minpts: number of mini points within circle). eps and minpts assign by user. Hybrid model help us to find the size of radius of the circle. Type of points include Core points(if minimum points available in its distance ), border points(end point where new circle not drawn or less than minimum points available in its area), outlier points(which don't match with any data point).

Done

Done

if data points are high and data points closely available then this algorithm can be efficient.

Lecture done