From the course: Machine Learning and AI Foundations: Clustering and Association

Unlock the full course today

Join today to access over 24,700 courses taught by industry experts.

How does HDBSCAN work?

How does HDBSCAN work?

- [Tutor] There's a more recently developed cluster analysis algorithm that's available now in Modeler called HDBSCAN. Now, hold on. That stands for hierarchical, density-based, spatial clustering of applications with noise. I want to break down the key points. We've heard hierarchical before, and it's going to be related to that. Also, it's density-based. That's opposed to centroid-based, which K-means was. And then also this idea of noise. All of those are going to be important. So let's compare it to K-means. First, rather than trying to identify a centroid, it identifies areas of high density. You see, the problem with centroids can be that they assume a spherical shape to the clusters. You'll sometimes hear experts refer to it as a Gaussian ball, which simply means if you've got a dozen dimensions, you're really not talking about a sphere anymore. But we can't assume that our data is shaped like that. And…

Contents