A comparison of agglomerative clustering methods with respect to noise
Janice L. DuBien, William D. Warde · Communication in Statistics- Theory and Methods · 1987
This article delineates the important considerations in any extensive, systematic comparison of clustering methods, and then it presents an empirical investigation of the effect of correlated variables on the “retrieval” ability of a particular class of agglomerative clustering methods. The empirical investigation re- sulted in two major observations; namely, that there are many agglomerative clustering methods, which have never been applied, that are “better” than any of the commonly used agglomerative clus- tering methods, and that correlated variables affect the “retrieval” ability of different agglomerative clustering methods differ- ently.