To understand what clustering is, let’s begin with an applicable example. Let’s say you want to travel to 20 places over a period of four days. How can you visit them all? We can come to a solution using clustering, and grouping the places into four sets (or clusters).
To determine these clusters, places that are nearest to one another are grouped together. The result is four clusters based on proximity, allowing you to visit all 20 places within your allotted four-day period.
Clustering is the method of dividing objects into sets that are similar, and dissimilar to the objects belonging to another set. There are two different types of clustering, each divisible into two subsets
- Hierarchical clustering
- Agglomerative
- Divisive
- Partial clustering
- K-means
- Fuzzy c-means
Every kind of clustering has its own purpose and numerous use cases.
Customer Segmentation
In customer segmentation, clustering can help answer the questions:
- What people belong to together?
- How do we group them together?
Social Network Analysis
User personas are a good use of clustering for social networking analysis. We can look for similarities between people and group them accordingly.
City Planning
Clustering is popular in the realm of city planning. Planners need to check that an industrial zone isn’t near a residential area, or that a commercial zone somehow wound up in the middle of an industrial zone.
However, in this article, we’ll focus on hierarchical clustering.