Question

State similarities and differences between Fuzzy c-means and hierarchical clustering based on Gaussian distributions.

Answer 1

A) Fuzzy-C means is a clustering method based on an iterative methodology that assigns a set of discrete (Boolean) class membership values on the basis of the distance in feature space between a feature vector and each class centroid.

Hlerarchlcal based clustering-> In this method of clustering there is a nested structure such that within each cluster there is again a cluster, such that they are overlapping. In the tree structure, the root being th cluster containing everything. The parent is the union of its children and the leaf nodes are one subcluster 1. Single level partioning partitions into a constant number of clusters (that has to be specified prior to processing )where as hierarchical partitions into unknown numbers of partition which may vary. 2. The advaltage of Single level partitining is it is very effecient in terms of time, It taskes less time than hierarchical clustering. 3. Heirarchical clustering always produces the same result and its consistent, where as Single level partioning may produce different result because it randomly choses the clusters centers, so result may vary 4. Heirarchical is good if our data set is small, where as Single level partioning is good for large data sets. s.wiehe rfmance of her wbthe ime talen to execute increases. 6. Hiererchical clustering works well when the underlying data is hiererchical where as for single level partioning it works well if the data is spherical or we can say circular.