DBSCAN
DBSCAN is another clustering quite popular technique. DBSCAN belongs to the density-based family of algorithms, and, unlike K-Means, it does not need to know the number of clusters, K, in advance.
In a few words, DBSCAN works as follows: at each step it takes an item to grow a cluster around it.
When we take an item from a high-density region, then there are many other data points close to the current item, and all these items are added to the cluster. Then the process is repeated for each newly added element of the cluster. If, however, the region is not dense enough, and there are not so many points nearby, then we do not form a cluster and say that this item is an outlier.
So, for DBSCAN to work, we need to provide the following ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access