Goal of this project is to find the best algorithm to represent given clusters.
My input is a number of clusters in any space that has a similarity measure.
The first part of the project, finds the best representation of given cluster. Different algorithms will be tried:
- random selection
- greedy aproach
- improved greedy approach
- k-means
- more?
When I have the representation I will test whether given the points from that cluster will be assigned by kNN algorithm to the right cluster. Therefore I will be able to compare the algorithms above.