Maximize the similarity between different clusters

Minimize the similarity within each cluster

Maximize the similarity within each cluster

Minimize the number of clusters

It can handle overlapping clusters

It requires the number of clusters to be known beforehand

It uses hierarchical methods for clustering

It is not sensitive to initial cluster centers

Actual data points (medoids)

Centroids calculated from the data

Mean of the data points

Randomly selected points

Random sampling

Silhouette analysis

Principal component analysis

Hierarchical clustering

It can handle any shape of clusters

It is sensitive to outliers

It requires a large amount of data

It is computationally intensive

Assign points to the nearest centroid

Calculate the distance metric

Choose the number of clusters (k)

Update centroids

To visualize the clusters formed

To determine the optimal number of clusters

To calculate the distance between points

To initialize the centroids