In this case of clustering, the hierarchical decomposition is done with the help of bottom-up strategy where it starts by creating atomic (small) clusters by adding one data object at a time and then merges them together to form a big cluster at the end, where this cluster meets all the termination conditions. This procedure is iterative until all the data points are brought under one single big cluster.
Basic algorithm of agglomerative clustering
- Determine the proximity matrix.
- Assume that each data point belongs to a cluster.
- Do it again.
- Combine the two groups that are the closest together.
- Make changes to the proximity matrix.
- Continue until just one cluster remains.
0 comments:
Post a Comment
Let us know your responses and feedback