This program is an implementation of K-means / K-medoids Clustering algorithms in language C++. The program implements algorithms for vector clustering in d-dimensional space, using 12 combinations of the following variations. Euclidean and Cosine distances are used.
Initialization
1.Random selection of k points (simplest)
2.K-means++
Assignment
1.Lloyd’s assignment
2.Assignment by Range search with LSH
3.Assignment by Range search with Hypercube
Update
1.K-means
2.Partitioning Around Medoids (PAM) improved like Lloyd’s
How to run(using terminal):
./cluster -i <input file> -c <configuration file> -o <output file> -d <metric>