ProClus: The ProClus Algorithm for Projected Clustering

In subspace: Interface to OpenSubspace

Description

The ProClus algorithm works in a manner similar to K-Medoids. Initially, a set of medoids of a size that is proportional to k is chosen. Then medoids that are likely to be outliers or are part of a cluster that is better represented by another medoid are removed until k medoids are left. Clusters are then assumed to be around these medoids.

Usage

ProClus(data, k = 4, d = 3)

Arguments

data	A Matrix of input data.
k	Number of Clusters to be found.
d	Average number of dimensions in which the clusters reside

References

C. C. Aggarwal and C. Procopiuc Fast Algorithms for Projected Clustering. In Proc. ACM SIGMOD 1999.

See Also

Other subspace.clustering.algorithms: CLIQUE; FIRES; P3C; SubClu

Examples

data("subspace_dataset")
ProClus(subspace_dataset,k=12,d=2.5)

Example output

OpenJDK 64-Bit Server VM warning: Can't detect initial thread stack location - find_vma failed
Subspace clustering generated by the package Subspace, containing 12 clusters.Warning message:
system call failed: Cannot allocate memory 

subspace documentation built on May 2, 2019, 11:11 a.m.