Optimal Variable WeightingA useful tool for ultrametric and additive tree clustering | |
Download |
Optimal Variable Weighting Ranking & Summary
Advertisement
Optimal Variable Weighting Tags
Optimal Variable Weighting Description
The OVW application performs optimal variable weighting for ultrametric and additive tree clustering as well as for K-means partitioning, following the method proposed by De Soete (1986, 1988), and Makarenkov and Legendre (2001). Given a rectangular data matrix Y (n,m), containing measurements of n objects on m variables, the procedure computes variable weights w(m) such that the resulting matrix of inter-object dissimilarities D(n,n) obtained from Y optimally satisfies either the ultrametric or the additive inequality, or optimally corresponds to a K-means partition with fixed number of groups K. The weights w are constrained to be nonnegative and their sum is equal to one. Take Optimal Variable Weighting for a test drive to fully assess its capabilities!
Optimal Variable Weighting Related Software