A New Dynamic Programing Algorithm for Cluster Analysis
Thierry Gafner · Computational Statistics · 1992
With this study we showed in a first phase using intensive simulations, that only the mathematical programing methods could find an optimal solution to to a cluster analysis problem; and that the classical methods only could find an approximation. In a second phase we tried to reduce the complexity of the dynamic programing algorithm. To do that, we combined elements taken in three methods.