Original article was published by Manik Soni on Artificial Intelligence on Medium

# What is Hierarchical Clustering?

## What is Hierarchical Clustering? How does Hierarchical Clustering work? What is meant by Hierarchical Clustering? Types of Hierarchical Clustering.

**Hierarchical Clustering** may be defined as a way to group similar objects into clusters. So it requires the whole concept of successively merging all the clusters into one single cluster. There are 2 types of hierarchical clustering agglomerative and divisive.

## Types of Hierarchical Clustering :

**Agglomerative Hierarchical Clustering:**It is a type of hierarchical clustering which uses a bottom-up approach to make clusters. It uses an approach of the partitioning of 2 most similar clusters and repeats this step until there is only one cluster.**Divisive Hierarchical Clustering:**It is a type of hierarchical clustering that uses a top-down approach to make clusters. It uses an approach of the partitioning of 2 least similar clusters and repeats this step until there is only one cluster.

## How does Hierarchical Clustering work?

There are steps that help to form Hierarchical Clusters.

**Step 1.** Make each data point a single-point cluster, that forms ’N’ number of clusters.

**Step 2. **Take the two closest data points and make them one cluster so for that, we are using **Euclidean Distance**.

Now we in order to take the closest clusters we are using 4 options.

**Option 1. **Take the distance between** Closest Points.**

**Option 2.** Take the distance between** Furthest Points.**

**Option 3.** Take the distance between** Average Distance **that is take the average of all the distance between the points and take the closest one**.**

**Option 4.** Take the distance between** Centroids.**

Now, back to **Step 2 **applying this distance formula and choose one of the above approaches.

**Step 3.**Take the 2 closest clusters and make them one cluster, which is 4 clusters.

**Step 4.** Repeat **Step 3** until there is only one cluster.

Again Repeat the steps.

Now your model is ready when all the clusters are combined to one single cluster.