A Method of Using Cluster Analysis to Study Statistical Dependence in Multivariate Data
William J. Borucki, Don H. Card, G. C. Lyle · IEEE Transactions on Computers · 1975
A technique is presented that uses both cluster analysis and a Monte Carlo significance test of clusters to discover associations between variables in multidimensional data. The method is applied to an example of a noisy function in three-dimensional space, to a sample from a mixture of three bivariate normal distributions, and to the well-known Fisher's Iris data.