Inclusion of irrelevant variables in a cluster analysis adversely affects subgroup recovery. This paper examines using moment-based statistics to screen variables; only variables that pass the screening are then used in clustering. Normal mixtures are analytically shown often to possess negative kurtosis. Two related measures, "m" and…

Two Monte Carlo studies investigated the effects of within-group covariance structure on subgroup recovery by 10 hierarchical clustering methods using 100 bivariate observations from 2 subgroups. Superior recovery was associated with within-group correlation that matched the direction of subgroup separation. (SLD)

This article examines using moment-based statistics to screen variables that are then used in clustering. A Monte Carlo study found that screening variables was a viable alternative to both ultrametric weighting and forward selection of variables. Advantages and disadvantages of screening are discussed. (SLD)

Used Monte Carlo methods to evaluate the adequacy of cluster analysis to recover group membership based on simulated latent growth curve (LCG) models. Cluster analysis failed to recover growth subtypes adequately when the difference between growth curves was shape only. Discusses circumstances under which it was more successful. (SLD)

Presents a variable selection heuristic for nonhierarchical (K-means) cluster analysis based on the adjusted Rand index for measuring cluster recovery. Subjected the heuristic to Monte Carlo testing across more than 2,200 datasets. Results indicate that the heuristic is extremely effective at eliminating masking variables. (SLD)

One set of approaches to the problem of clustering with dichotomous data in cluster analysis (CA) was studied. The techniques developed for clustering with binary data involve calculating distances between observations based on the variables and then applying one of the standard CA algorithms to these distances. One of the groups of distances that…

Evaluated the performance of several alternative cluster analytic approaches to initial model specification using population parameter analyses and a Monte Carlo simulation. Of the six cluster approaches evaluated, the one using the correlations of item correlations as a proximity metric and average linking as a clustering algorithm performed the…

A simulation study compares nonhierarchical clustering capabilities of a class of neural networks using Kohonen learning with a K-means clustering procedure. The focus is on the ability of the procedures to recover correctly the known cluster structure in the data. Advantages and disadvantages of the procedures are reviewed. (SLD)

A Monte Carlo study compared the usefulness of six variable weighting methods for cluster analysis. Data were 100 bivariate observations from 2 subgroups, generated according to a finite normal mixture model. Subgroup size, within-group correlation, within-group variance, and distance between subgroup centroids were manipulated. Of the clustering…

A Monte Carlo study, involving 6,000 "computer subjects" and three raters, explored the reliability of the inverse screen test for cluster analysis. Results indicate that the inverse screen may be a useful and reliable cluster analytic technique for determining the number of true groups. (TJH)

A Monte Carlo study evaluated the effectiveness of three rules of classifying objects into clusters: nearest neighbor classification; nearest centroid assignment; and quadratic discriminant analysis. Results suggest that the nearest neighbor rule is a useful tool for assessing the validity of the clustering procedure of J. H. Ward (1963). (SLD)

The partitioning of squared Euclidean--E(sup 2)--distance between two vectors in M-dimensional space into the sum of squared lengths of vectors in mutually orthogonal subspaces is discussed. Applications to specific cluster analysis problems are provided (i.e., to design Monte Carlo studies for performance comparisons of several clustering methods…

Monte Carlo studies investigated effects of within-group covariance structure on subgroup recovery by several widely used hierarchical clustering methods. In Study 1, subgroup size, within-group correlation, within-group variance, and distance between subgroup centroids were manipulated. All clustering methods were strongly affected by…

A method is presented that simultaneously estimates cluster membership and corresponding regression functions for a sample of observations or subjects. This methodology is presented with the simulated annealing-based algorithm. A set of Monte Carlo analyses is included to demonstrate the performance of the algorithm. (SLD)

A Monte Carlo study of the validity of the Inverse Scree Test under conditions where true group membership is known was conducted. Fifty cluster analyses of each distribution involving 2 to 5 true groups of 3,000 simulated subjects were made. Implications for the data analyst are discussed. (SLD)