Partition-Based Clustering

The overall strategy here is to partition the Dataset into a set of $k$ clusters $C_i$ such that the sum of squared distances to $c_i$ is minimized, where $c_i$ is the centroid or medoid of a cluster $C_i$. $\min {\sum }_{i=1}^k({\sum }_{j=1}^{n}d{\left(o_j,c_i\right)}^2$ You could enumerate all possible partitions and therefore find the optimal solution or you could use heuristic methods like:

• k-means Clustering
• k-medoids Clustering
• Clustering Large Applications
• Randomized CLARA