Legend
ABJ Normal
ABJ Tidak Normal

Sesuaikan Klaster

Hasil klaster terlihat pada peta diatas. Anda dapat menyesuaikan klaster dengan mengubah nilai epsilon dan min points.

Cluster Analysis: DBSCAN

Data clustering is a fundamental task in machine learning and data analysis. One powerful technique that has gained prominence is Density-Based Spatial Clustering of Applications with Noise (DBSCAN). In this blog, we delve into the world of DBSCAN, exploring its principles and applications in uncovering hidden structures within datasets.

Join us on a journey to understand how DBSCAN goes beyond traditional clustering methods, offering a unique approach to identifying clusters based on the density of data points. Let’s unravel the intricacies of DBSCAN and unlock its potential for unraveling patterns in your data.

Idea behind density based clustering :

  1. Density based clustering algorithms divides your entire dataset into dense regions separated by sparse regions.

MinPts and Epsilon :

  1. Measuring density around a point is straightforward — we define a region around the point and assess the number of points within that designated area. This approach serves as a practical method for gauging the density surrounding a specific point.
  2. To determine density around a point, we employ circles in 2-D, spheres in 3-D, and hyper-spheres in n-dimensional spaces. Suppose we draw unit radius circle around a point P as shown in above figure and here we establish a criterion: a region is considered sparse if it contains fewer than 3 points and dense if it contains 3 or more points.
  3. MinPts stands for “Minimum Points”, is a parameter that specifies the minimum number of points required to form a dense region, which is consider a cluster.

Core Points, Border Points and Noise Points :

A point is considered a core point if it has a minimum number of other points(specified by MinPts) within a given radius ε of itself.

In the depicted diagram, with ε set to 1 and MinPts to 4, let’s focus on a specific point, P. To determine if P qualifies as a core point, we create a circle with a radius of 1 unit around P. Observing the diagram, it’s evident that point P, along with three additional points within the circle, satisfies the MinPts condition. Hence, we can confidently classify point P as a core point.

Examining the diagram, it’s evident that within the circle surrounding a specific point, there are only two points in addition to the point itself, totaling three points. This doesn’t meet the MinPts requirement of 4, leading us to conclude that it is not a core point.